Proceedings of the International Workshop on EuroPLOT Persuasive Technology for Learning, Education...

IWEPLET 2013

The EuroPLOT project and the IWEPLET 2013 workshop are funded by the Education,

Audiovisual and Culture Executive Agency (EACEA) of the European Commission through the Lifelong Learning Programme (LLL) by grant #511633 from 2010 – 2013.

ii

The EuroPLOT Project Partners:

Aalborg University (Denmark) hosts the Centre for Persuasive Design, is the originator of the EMDROS

database and the 3ET tool. Aalborg University also provides target groups of teachers in archival work, language learning and environmental science.

DHI (Denmark) is an international consulting and research organisation, and brings expertise in the development of technical guidance and courses for vocational training, including the application of e- learning resources and tools, and represents the target group on environmental science.

The University of Hradec Králové (Czech Republic) has led several European projects on learning objects, particularly on their creation and distribution through repositories (for example e-dilema), and is experienced in using e-learning in teaching business computing.

Danube University Krems (Austria) has a large interest in Open Educational Repositories (OER), the development of e-learning taxonomies and learner- centred design approaches related to learning environments, and has significant experience of evaluating educational projects and expertise in learning design and pedagogic frameworks.

Leeds Metropolitan University (UK) is the lead partner on this project and has expertise in developing learning objects (participation in 2 European projects), practical experience in Open Educational Repositories (OER) (institutional OER and repository projects), expertise in learning design patterns (led a national project), and in natural language processing and artificial intelligence applied to education (national project).

London Metropolitan University (UK) brings experience from the Learning Technology Research Institute with an international reputation in research on learning objects and learning design, and they are developers of GLOmaker, a generative learning object authoring tool.

iii

Proceedings of the

International Workshop on EuroPLOT Persuasive Technology for Learning,

Education, and Teaching

(IWEPLET 2013)

16-17 September 2013

Paphos (Cyprus)

Editors: Reinhold Behringer and Georgina Sinclair (Leeds Metropolitan University)

IWEPLET 2013 is held in conjunction with the EC-TEL 2013 conference

4

The papers in this book comprise the proceedings of the meeting mentioned on the cover and title

page. They reflect the authors’ opinion and, in the interest of timely dissemination, are published as

presented and without change. Their inclusion in this publication does not necessarily constitute

endorsement by the editors.

(2nd edition, with ninor revisions)

5

Preface

This book contains the proceedings of the International Workshop on EuroPLOT Persuasive Technology for Learning, Education and Teaching (IWEPLET) 2013 which was held on 16-17 September 2013 in Paphos (Cyprus) in conjunction with the EC-TEL conference. The workshop and hence the proceedings are divided in two parts: on Day 1 the EuroPLOT project and its results are introduced, with papers about the specific case studies and their evaluation. On Day 2, peer-reviewed papers are presented which address specific topics and issues going beyond the EuroPLOT scope.

This workshop is one of the deliverables (D 2.6) of the EuroPLOT project, which has been funded from November 2010 – October 2013 by the Education, Audiovisual and Culture Executive Agency (EACEA) of the European Commission through the Lifelong Learning Programme (LLL) by grant #511633. The purpose of this project was to develop and evaluate Persuasive Learning Objects and Technologies (PLOTS), based on the ideas of BJ Fogg. The purpose of this workshop is to summarize the findings obtained during this project and disseminate them to an interested audience. Furthermore, it shall foster discussions about the future of persuasive technology and design in the context of learning, education and teaching.

The international community working in this area of research is relatively small. Nevertheless, we have received a number of high-quality submissions submitted to a peer- review process before being selected for presentation and publication. We hope that the information found in this book is useful to the reader and that more interest in this novel approach of persuasive design for teaching/education/learning is stimulated.

We are very grateful to the organisers of EC-TEL 2013 for allowing us to host IWEPLET 2013 within their organisational facility which was a great help in preparing this event.

I am also very grateful to everyone in the EuroPLOT team for collaborating so effectively during the three years towards creating excellent outputs, and for being such a nice group with a very positive spirit also beyond work.

And finally I would like to thank the EACEA for providing the financial resources for the EuroPLOT project and for being very helpful when needed. This funding made it possible to organise the IWEPLET workshop without charging a fee from the participants.

Reinhold Behringer

Chair of IWEPLET 2013

6

Committee and Chairs

Programme Chair:

Janet Finlay, Cardiff University

Programme Committee:

Peter Baumgartner, Danube University Krems Tom Boyle, London Metropolitan University Sandra Burri Gram-Hansen, Aalborg University Pavel Čech, University of Hradec Králové John Elliott, Leeds Metropolitan University Christian Grund Sørensen, Aalborg University Elizabeth Guest, Leeds Metropolitan University Erich Herber, Danube University Krems Mikulecká Jaroslava, University of Hradec Králové Peter Øhrstrøm, Aalborg University Carl Smith, London Metropolitan University Margrethe Winther-Nielsen, DHI Nicolai Winther-Nielsen, Aalborg University

Workshop Management Chair:

Georgina Sinclair, Leeds Metropolitan University

General Chair:

Reinhold Behringer, Leeds Metropolitan University

vii

Table of Contents

Part A: The EuroPLOT Project

Persuasive Technology for Learning and Teaching – The EuroPLOT Project 3 Reinhold Behringer, Mekala Soosay. Sandra Burri Gram-Hansen, Peter Øhrstrøm, Christian Grund-Sørensen, Carl Smith, Jaroslava Mikulecká, Nicolai Winther-Nielsen Margrethe Winther-Nielsen, Erich Herber

PLOTMaker – Persuasive Learning Design through Context Engineering 9 Carl Smith and Valentina Chinnici

PLOTLearner as Persuasive Technology: Tool, Simulation and Virtual World for Language 21 Learning

Nicolai Winther-Nielsen

Case Study: Chemical Handling – PLOTMaker for Training in Exposure Scenarios under 29 REACH

Margrethe Winther-Nielsen and Ilaria De Rosa Carstensen

Case Study: Applying Persuasive Principles to Influence Students in Adopting Deeper Learning 37 Approaches

Mekala Soosay and Jaroslava Mikulecká

Case Study: Kaj Munk: Using Persuasive Learning 45 Christian Grund-Sørensen, Sandra Burri Gram-Hansen, and Peter Øhrstrøm

Case Study: PLOTLearner for a Corpus of the Hebrew Bible: The Case for Repurposing in 53 Language Learning


Evaluation: Designing the Persuasive Learning Experience 61 Erich Herber

Part B: Beyond EuroPLOT

Persuasive Design 69

Persuasive Design – A Matter of Context Adaptation? 71 Sandra Burri Gram-Hansen

Persuasive Design of TEL: Models and Challenges 79 Kristian Torning

Immersive Layers Design Exploring Culture through a Persuasive Multimodal Interface 91 Christian Grund-Sørensen

A Unified Framework for Analysing, Designing and Evaluating Persuasive Technologies 99 Isaac Wiafe

viii

Assessment and Learning Success 107

Persuasive Skill Development: On Computational Surveillance Strategies for Modeling Learning Statistics with PLOTLearner

Judith Gottschalk and Nicolai Winther-Nielsen

109

A Set of Criteria to Assess Motivation and Persuasion of e-Learning Applications 117 Eric Brangier and Michel Desmarais

Specific Technologies for Persuasive Learning 125

Augmented Reality in Persuasive Learning: The Kaj Munk Case 127 Casper Justesen Bech

Intuitive Surveying & Quantification of Qualitative Input through the Conceptual Pond 135 Christian Grund-Sørensen

Architecture of Applications Built on Emdros: Case Studies in Systems for Persuasive Learning 143

Ulrik Sandborg-Petersen

Application Domains 151

Motivating the Interest in Danish Literature with Mobile Persuasive Learning 153 Sandra Burri Gram-Hansen, Karina Dyrby Kristensen and Lasse Burri Gram-Hansen

PEERs at Play: A Case Study on Persuasive Educational and Entertainment Robotics in Autism Therapy and Education

Lykke Brogaard Bertel and Dorte Malig Rasmussen

161

Part A: The EuroPLOT Project

IWEPLET 2013 Proceedings - Page 1 of 168

Persuasive Technology for Learning and Teaching - The EuroPLOT Project

Reinhold Behringer, Mekala Soosay

Leeds Metropolitan University, Leeds LS6 3QS, United Kingdom

{r.behringer, m.soosay}@leedsmet.ac.uk

Sandra Burri Gram-Hansen, Peter Øhrstrøm, Christian Grund Sørensen

Aalborg University, Denmark

{burri, poe, cgrund}@hum.aau.dk

Carl Smith

London Metropolitan University, London, United Kingdom

[email protected]

Jaroslava Mikulecká

University of Hradec Králové, Czech Republic

[email protected]


Fjellhaug International University College, Denmark

[email protected]

Margrethe Winther-Nielsen

DHI

[email protected]

Erich Herber

Danube University Krems

[email protected]

The concept of persuasive design has demonstrated its benefits by changing human

behavior in certain situations, but in the area of education and learning, this approach has

rarely been used. To change this and to study the feasibility of persuasive technology in

teaching and learning, the EuroPLOT project (PLOT = Persuasive Learning Objects and

Technologies) has been funded 2010-2013 by the Education, Audiovisual and Culture

Executive Agency (EACEA) in the Life-long Learning (LLL) programme. In this

program two tools have been developed (PLOTMaker and PLOTLearner) which allow

the creation of learning objects with inherently persuasive concepts embedded. These

tools and the learning objects have been evaluated in four case studies: Language

Learning (Ancient Hebrew), Museum Learning (Kaj Munk Museum, Denmark),

Chemical Handling, and Academic Business Computing. These case studies cover a

wide range of different learning styles and learning groups, and the results obtained

through the evaluation of these case studies show the range of successes of persuasive

learning. They also indicate the limitations and areas where improvements are required.

Keywords: Persuasive Design, persuasive technology, eLearning, learning objects, case

studies


Background In his seminal introduction of the concept of Persuasive Technology (PT), BJ Fogg [1] has

identified several core principles of persuasive design of interaction systems. These principles

have been shown to be able to change human behavior through indirect and informal

learning. The question arose as to whether these principles could also be put to use in more

formal learning. In 2010 the EuroPLOT consortium [2] was formed to address this problem

and to investigate the use of these persuasive principles in learning and teaching. In this

project, which is funded by the Education, Audiovisual and Culture Executive Agency

(EACEA) of the European Commission in the Lifelong Learning Programme from

2010-2013, we have developed a framework for developing persuasive learning designs [6].

This framework was used to inform the development of two tools which allow the creation of

persuasive learning objects based on persuasive learning designs: PLOTMaker and

PLOTLearner. These two tools have been developed from existing software which has been

extended with the persuasive elements and design options. These tools and the learning

objects created with them have been applied and evaluated in four different case studies with

groups of teachers and learners from realms with distinctly different teaching and learning

practices: Academic business computing, language learning, museum learning, and chemical

substance handling. Altogether, these case studies have involved the following wide range of

learner target groups: school children, university students, tertiary students, vocational

learners and adult learners. With regards to the learning context, they address archive-based

learning, industrial training, and academic teaching. The participants in these case studies are

from Sweden, Africa (Madagascar), Denmark, Czech Republic, and the UK.

Main elements of Persuasive Design The main idea of the EuroPLOT project is to develop technologies which would allow the

designers of learning objects to create such objects with inherently built-in “persuasiveness”.

The seven persuasive design principles identified by BJ Fogg are:

1 Reduction refers to the design strategy of simplifying what would otherwise be a complex

process. For example, Amazon’s 1-click purchase which allows you to skip a lot of

time consuming navigations and tedious form filling, in order to make an instant

2 Tpu

urnchnaeslien. g is a design strategy which places the user inside a process that has a pre-

determined direction. E.g. most installation processes require that the user completes

several steps before the installations process is completed.

3 Tailoring is the degree to which a site or a program presents relevant content to individual

users or user groups. Navigational options, filtering mechanisms and labelling systems can

all be adapted to reflect user demographics.

4 Suggestion is the persuasive design strategy of delivering a message at the opportune

moment. E.g. when Amazon suggests extra books which are closely related to the one you

were just about to buy. - It is very important that suggestions are made at the right time,

Kairos. This idea stems from ideas of Greek philosophers, in particular Aristotle. It may be

defined as the opportune moment to perform a persuasive action. It is a quite powerful

concept which is not easily formalised, but its use in technology applications can

significantly contribute to persuasiveness by the just-in-time and just-in-place paradigms

5 Self-monitoring is the design strategy which allows you to monitor progress. E.g. sites

which require a log-in and then enables the user to monitor the progress of weight loss.


6 Surveillance is closely related to self-monitoring; however the monitoring is not done by

the user but by the system or the owners of the system. E.g. when using a weight loss

website, users may be motivated not only by monitoring their own progress, but also by

sharing experience and receiving feedback from other users who are struggling with

similar issues. By sharing statistics, diet-plans etc. users feel more related to each other and

may be inspired by actions taken by others.

7 Conditioning refers to the strategy of embedding emotional feedback into a design. It is

often expressed as praise and rewards, but in a slightly more subtle manner than the case

with Persuasive Social Actors. E.g. when forums reward users with increasingly lofty

titles (or user rights) in correlation to the number of posts made by the user.

Tool development and case studies Tools

Two tools were developed in the course of the EuroPLOT project for creating such

persuasive learning objects: PLOTMaker stems from the software tool GLOMaker [3] which

allows the creation of self-contained digital learning objects and which has already been in

use and development for several years before the EuroPLOT project. The additions to convert

GLOMaker into PLOTMaker during the EuroPLOT project include several elements that

refer to the persuasive principles, which can be put into the learning objects created with

PLOTMaker. In particular the Augmented Reality component in PLOTMaker demonstrates

the Kairos principle of just-in-place for location-based learning.

The tool PLOTLearner has been derived from the tool 3ET [5] and uses a large corpus of

data in which language tests are stored, based on the database EMDROS [4]. This database is

annotated and allows retrieval of text, broken into components of the language structure

(words, phrases, clauses etc.). This very specific capability makes PLOTLearner suitable for

learning where a large corpus of text is available. Case studies

These two tools have been used to create learning objects in four case studies for

demonstrating the viability of the persuasive learning approach. Altogether, these case studies

cover several different learning styles, learner and teacher groups, countries and cultures. As

target learners they address school children, university students, tertiary students, vocational

learners and adult learners. With regard to the learning context, they address archive-based

learning, industrial training, and academic teaching. Internationally these case studies include

participants from Sweden, Africa (Madagascar), Denmark, Czech Republic, and UK. These

case studies do not claim to give a comprehensive answer to the question of persuasive

learning which does have a deeper positive impact than traditional learning, but show that

this approach does have benefits for learning success. The reason for involving such a

breadth of topics, learners, and learning styles was to demonstrate the applicability of

persuasive learning across a wide range.

Academic Business Computing

This case study is undertaken in the framework of database teaching in academic computer

courses which are taught at two universities. The learning objects in these courses are

developed with the tool PLOTMaker and focus on teaching basic SQL. The goal of this case

study is to demonstrate the applicability of this approach in two different countries and

languages (English in UK, Czech in Czech Republic). The specific topic of these learning

objects is database normalization and SQL querying, and the main persuasive principles used

in the PLOT’s design are reduction and interactivity.


Language Learning

This case study investigated the learning of language with the help of a large corpus of text.

The learning tool PLOTLearner was specifically developed for this type of learning with a

large text repository (data-driven learning) from annotated texts. The language to be taught is

Ancient Hebrew. Due to the difficulty of this language, this provides a valuable example of

how a language with a different visual writing system can be taught effectively to students

through the engagement with a large structured text corpus.

Mediating Kaj Munk

The Danish writer and vicar Kaj Munk (1898-1944) has produced a significant oeuvre of

plays and other texts, all of which are archived in the Kaj Munk Archive at Aalborg

University. In addition, there is a Kaj Munk Museum in his old Vicarage in Vedersø

(Denmark). Learners who want to inform themselves about life and work of this writer can do

so through several methods developed in the EuroPLOT project: they can explore the writings

through an online Kaj Munk Study Edition that allows structured access through the

EMDROS database of Kaj Munk’s works. Furthermore, a tool has been developed which

makes it possible to virtually visit the Kaj Munk Museum using mobile devices, thereby using

computers as persuasive media with Augmented Reality (AR) technology.

Chemical Substance Handling

In an industrial context, employees often need training in how to implement and handle new

regulatory demands. This is especially important in the case of using chemical substances

which may be harmful. In the EuroPLOT project, DHI has used PLOTMaker to develop

learning objects which teach adult learners in an industrial context how to handle such

dangerous chemicals. This considers health and safety aspects, and the persuasion of the

learning objects is implemented through tailoring and simulation.

Results The evaluation is in its final phase, and the results will be available and presented at EC-

TEL in September 2013. A common set of evaluation criteria was used for all case studies,

although this could not be consequentially maintained due to the fundamental differences of

the case scenarios.

The Business Computing case study uses an outcome-based learning design paradigm for

creating the courses, hosted on the virtual learning environment Blackboard Learn. Student

feedback was evaluated and overall positive attitude and reception was noted. Constructive

alignment has been found in the context of this case to be a key principle for which learning

objects can be perceived as persuasive and effective for learning.

The Language Learning case study for Ancient Hebrew has involved learners in Denmark,

Sweden and Madagascar. The tool PLOTLearner was successfully evaluated with 91 learners.

Instructional strategies that embed self-reflective learning activities and challenging

interactive exercises have been detected as particularly valuable components of a persuasive

design approach for learning the language.

The Kaj Munk case study has involved teachers and students in classroom and outdoors

(AR). One novel contribution was the Conceptual Pond in which personal reflections can be

collected in a simple and intuitive way.

The Environmental Science case study for teaching handling of harmful chemicals has

involved 22 participants. The process of creating learning objects did lead to further

improvements in the tool capabilities. In the context of this case study for adult education we


found that the use of explorative learning designs has shown potential towards creating an

awareness of the practical relevance of what was learned.

Future work The EuroPLOT project will conclude in October 2013. Workshops will be held to

disseminate the results of this project, including lessons-learned and the general

persuasive concepts for teaching and learning. Further work will be to conduct an actual

study on the true effectiveness of persuasive principles in teaching and learning, as these

case studies only provide an indication of the effectiveness of persuasive learning.

Acknowledgements. The EuroPLOT project was funded by the Education, Audiovisual and

Culture Executive Agency (EACEA) of the European Commission through the Lifelong

Learning Program with grant #511633.

References

1. Fogg, BJ: Persuasive Technology - Using Computers to Change What We Think and Do. 2003, San Francisco: Morgan

Kaufmann Publishers (2003)

2. EuroPLOT: Persuasive Learning Objects and Technologies. http://www.eplot.eu/ (2010)

3. GLOMAKER: http://www.glomaker.org/ (2013)

4. Emdros. Emdros - the database for analyzed and annotated text. http://emdros.org/ (2013)

5. PLOTLearner 2: http://www.ezer.dk/3ETusersguide/PL-2.0.1/en/intro.php (2012)


Persuasive Learning Design through Context Engineering

Carl Smith Learning Technology Research Institute (LTRI),

London Metropolitan University.

[email protected]

Valentina Chinnici Learning Technology Research Institute (LTRI),

London Metropolitan University.

Università degli Studi di Roma Tre,

[email protected]

EuroPlot has focused on extending PLOTMaker by adding new components, persuasive

learning patterns, a mobile player and learner journey manager. This new authoring

environment enables teachers, trainers and students in industry and educational

institutions to rapidly create pedagogically based contextual mobile learning apps. The

authoring environment structures these apps methodologically to allow for the

exploration of the inherent relationships between the learning content and the context of

that content.

Keywords: Mobile Learning, Persuasive Learning, PLOTMaker, Learning Context.

Introduction

PLOTMaker is designed to support the aims of the EuroPlot project by giving teachers,

trainers and students the tools necessary to create and adapt effective learning resources.

PLOTMaker has been extended on the basis that the persuasiveness of a learning design is

not dependent on the technology itself, but on how the technology is applied within a given

context:

1. PLOTMaker as context engineering tool: Recent research is beginning to show that fluidity

of thinking and learning relates to fluidity of movement. As a result of this persuasive

principle the new version of PLOTMaker enables teachers, trainers and students to

dynamically manipulate elements of the learning context. This is achieved by exploring the

generation of 'meaning' in the physical environment and the inherent relationship between

learning content and the context of that content.

2. PLOTMaker as mobile app maker: The new version of PLOTMaker enables teachers,

trainers and students to rapidly create pedagogically based mobile learning apps in an

intuitive and easy way. The persuasive power of this development will be explored with an

example from the Kai Munk case study (figure 1+2) where students will capture their

learning experience in situ using the mobile PLOTMaker player and then feed that content

back into the PLOTMaker environment for novel reflection and interrogation of their learning.

Background

PLOTMaker has a wide international user base with 18,198 visitors having made 27,331

visits to the website in 2012 alone and from 129 different countries. There have also been

5,024 PLOTMaker downloads in the same time period.


The WYSIWYG authoring tool provides powerful features with an easy-to-use interface. It is

open source and free for educational use. The PLOTMaker tool currently works on computers,

tablets and mobile phones and allows users to overlay digital sight, sounds and interactions

onto the physical world in order to create rich immersive and interactive experiences. Users

equipped with a mobile device running the PLOTMaker mobile player can move through the

physical world and trigger digital media interventions in response to physical events such as

location, proximity, time and movement.

In addition to being used to create and then repurpose learning resources in four subject areas

within the EuroPlot project (archival studies, language learning, environmental science and

business computing) it has been used to develop and deliver a wide range of learning

resources in a number of projects funded by bodies such as JISC, FDF (Foundation Degrees

Forward) and many UK Universities.

Case Studies

1. Kai Munk

Figure 1: EuroPlot Kai Munk case study

PLOTMaker is being used to develop resources for problem-based learning for adult learners

and school children around the Kaj Munk archive and will work with the Kaj Munk Research

Center and the Kaj Munk Museum in Vedersø.


Figure 2: The Kai Munk Learner Journey Interface

2. Rylands Greek papyrus

(http://www.glomaker.org/samples/Papyrus12/GLO_Player.html) This example PLOT

focuses on Rylands Greek papyrus and uses the PLOT access views component (see figure 3).

The access views component is effectively a tool within a tool that is used (in this instance) to

explore a variety of interpretations of a single artefact from three different disciplinary

perspectives (History, Papyrology - study of papyrus documents, and Museum Studies). The

objective is to promote an appreciation of the insights about the item and its wider historical

context by considering an object from more than one disciplinary perspective.

Figure 3: Rylands Greek papyrus Access views component.


3. Ravensbourne Induction Trail

The Ravensbourne Induction Trail is an interactive location-based media experience designed

by students for (the new intake of) students at Ravensbourne University. This process of co-

creation was encouraged through the PLOTMaker architecture - ice-breakering and team

building activities were triggered via the PLOTs in order to combine personal with

environmental experiences.

Using the PLOTMaker tool, and consequently the PLOTmaker player, the learners and their

teams are guided into different reflective challenges and are asked to re-conceptualize them

by recording short videos expressing their point of view, their strengths as a team, and show

their decision making and problem solving skills. The Induction Trail consists of six PLOTs,

each of which is connected to an area of the building and specific topic and activity.

The learners have to discover the environment and become an active part in the process of

knowledge construction. The entire experience is learner-centered. As a result the members

of this “Community of Practice” (Wenger, 2006) actually (co)produce the resources which

can in turn usefully be shared with other communities (Ryberg, 2008)

According to pedagogical constructivism the students enhance their personal potential

developmental level through their independent problem solving activities. Rather than taking

place under adult guidance these activities are actually mentored by the tasks suggested by

the PLOT, and supported through collaboration with each other (Vygotsky, 1978).

Following the simple activities designed for the learner, this process of technology-based

learning becomes a social process of collaborative knowledge building (Brown & Campione,

1994; Lave, 1991; Pea, 1993a; Scardamalia & Bereiter, 1996)1 for the creation of cultural

artefacts (Stahl, 2005) and learner creativity. During the induction trail the learners will use

not only the PLOTmaker player, but they will be encouraged to use other “social” platform,

such as Instagram and Twitter to generate their personal content and challenge other teams.

1. Scavenger hunt: photo challenge

The “Scavenger hunt: photo challenge” (see figure 4) is an ice-breaker and team building activity which uses the PLOT as a guide for understanding the different

courses offered at Ravensbourne University. Instead of the traditional “scavenger

hunt” they have to use their devices socially, tagging their content via Twitter.

All the content is then available not only for the team that created it, but also for all

the other members of the other teams and in general for all the Twitter and Instagram

users. The objective is to create a new form of knowledge based on the personal

experience of different learners, that spreads over the social networks through creative

activities, in a collaborative social environment.


Figure 4: Scavenger Hunt: photo challenge

2. Central Loan Resources

The “Central Loan Resources” is the place where the equipment of Ravensbourne

university is borrowed. In this case we have created a memory game (see figure 5) to

let the students familiarise with all distinctive equipment from different fields of

study. In order to create a more interactive experience we have used the 360 degrees

panorama feature. The students could be able to interact actively with the

environment and discover the equipment moving into the room (see figure 6). For this

activity the social use of Instagram and Twitter is required.

Figure 5: CLR activity explained


Figure 6: CLR 360 degrees panorama

4. Urban Education and Language Learning.

The final case study focuses on the design, implementation and evaluation of a location based,

context aware system for urban education. The initial design was iteratively adapted using

evolutionary prototyping for language learning.

The Urban Education tour is based on a small area known as Eden Grove close to London

Metropolitan University and it explores how schools from 1850 to the present day are

signifiers of both urban change and continuity of educational policy and practice. The

theoretical perspective was to allow collaborating learners to interact: with each other, with

the mobile phones and with the physical environment in order to generate their own context

for development within a Zone of Proximal Development.

Iteration 1: Urban Education

The tour was developed using the initial design outlined above in conjunction with a complex

interplay between mobile learning technologies, iconic physical infrastructures and

educational discourses. The aim was to visualise urban education through various collective

images and representations (cf: Durkheim’s notion of the social imaginaire). This was

intended to enable researchers in both formal and informal learning contexts, when combined

with the real, to examine past and present representations of urban form, and relate these to a

contemporary Iteration of urban education (Pratt-Adams, Maguire and Burn, 2010). The

overall intention was to create a digital ‘technoscape’ (Appadurai, 1996; Urry, 2006) to

represent urban land, space, and subjects using a combination of social and cultural scripts.

These included oral histories, local historical stories, and material elements that detail

changes in the urban form, such as old photographs of pupils, school buildings and historical

maps. The aim was for the researcher to move through the re-constructed landscape and thus

“perform that landscape” (Sheller and Urry 2006, p. 9).

The development and production process involved the following elements:


• Initial field work and documentation of the site;

• Capture and digitisation of oral histories, Pathe news clips and local historical stories;

• Capture and digitisation of material elements that detail changes in the urban form,

such as photographs depicting the evolution of school buildings and historical maps;

• PLOT production to support the underlying pedagogy of the tour.

The cameras on the smart phones were used to allow students to produce video podcasts of

themselves and take photos. This instant capture of report writing and note making in situ

was designed to promote real time reflection. Images captured with the phones were

automatically geo-spatially tagged with their location information using GPS. These smart

phones are also capable of instant upload of data to sites like Flickr. Finally QTVR

(quicktime VR) movies of the interiors of the structures under investigation can be viewed

and manipulated in real time on location (as access to the interiors was not available during

the fieldwork).

Users running the PLOTMaker player on a mobile device can move through the physical

world and trigger digital media with GPS via an invisible interactive map. Zones are initially

set up on a map which has been geo-referenced to the physical site. The training starts with

minimal instruction for the user. The intention is that whilst the technology (GPS) is working

behind the scenes the content is very much at the forefront in order to minimise any technical

concerns. Once they have entered a zone, audio and textual instructions are automatically

triggered to the mobile device.

Some examples of the varied learning activities involved in the application include a section

where the user is asked to examine both the physical architecture and the virtual architecture

in the same physical location. The virtual architecture in this instance includes areas which

were not available to view on the day of the tour and visualizations of the building as it was

in the late 19th century. The user is then asked to examine what the building was originally

used for when it was established in 1870. The user also has the opportunity to listen to the

oral history of a former pupil at the school and adopt their point of view whilst in the same

physical space where the events took place. The user can reinvest the insight gained back into

the context and augment the space.

In another section the user is asked to look at a newsreel of a religious procession from the

1930s that was filmed in Eden Grove whilst they are standing in the same location where the

film was shot. The students can reflect on the significance of religion (in this case Roman

Catholicism) on the locale and its influence on schooling.

In the 1920s this area was known as the Ring Cross Estate and was in the second highest

criterion for overcrowding and squalor, with people living in some of the worst slums. During

the middle of the 20th Century, the area was part of a slum clearance programme. Conditions

improved throughout the 1970s. The user is shown these street scenes and asked to

approximate by physically sketching out how much of the area in the archive footage still

remains and how much of it has been redesigned. The student is encouraged to reflect on the

impact of the social conditions on educational standards.

The final section exposes the user to the differing architectural styles of the buildings. The

user is introduced to a traditional Victorian 3-decker style school design (where the hall is

located at the centre with classrooms coming off a central point) and asked to compare it to

the more recent open style designs where each classroom is given some access to the

playground. The central activity is to examine what the architecture suggests about the

educational approaches of the time.


Iteration 2: Language Learning

The aim of this iteration of the design was the realisation from the previous instances shown

above that the way information is spatially represented directly impacts on our cognition. “In

the everyday world, humans organise and manipulate objects in space to facilitate thinking.

We are constantly organising and reorganising space to enhance performance” (Kirsch,

1995). All forms of media use some level of spatial arrangement to organise information.

Computer-assisted language learning (CALL) has embraced virtual learning environments

such as Second Life but using actual physical structures as a means of scaffolding language

learning has not yet been explored. The method of loci was adopted by ancient orators in

order to remember and organise speeches using a combination of visual memory and location

(Yates, 92).

“The problem of seeing and retaining complex information is older than print…. The

principle ancient mnemonic device was called ‘The Method of Loci’ and places its

emphasis on memorability (via intelligibility and transparency) through visual

structures such as concept maps. It is very much concerned with the acquisition of new

knowledge. It plays upon methods that we use informally, and it is a tradition that

survives today …..Visualization was in itself an important method of theory building. It

may have been more important than text.” (Won, J. Storkerson, P. 1996)

So would our pre-existing spatial design act as a catalyst and activate the abilities of language

learners if carried out in situ? The tour was translated into four languages: German, Spanish,

Italian and French but the content remained exactly the same. The users were tested in the

classroom using this content as text only and then in situ using our design.

Although we accept that our description above is brief, we point out that reusing another

iteration of our design here for language learning was very easy and quick to achieve,

something that we see as a pointer to the generality of our approach. In the future research

section we will explore ideas for extending the Language Learning design.

Social Media

A potential way of extending the urban education project is to incorporate the dynamic use of

real time social data. We become part of a larger social identity through passing around ideas.

We spread ideas around through dialogue and other forms of interaction. They become

‘contagious patterns of cultural information that pass from meme to meme which in turn have

the ability to change the actions of a group’ (Dawkins, 1976). A meme is a basic unit of

cultural ideas, social semiotic symbols or practices, which can be transmitted from one mind

to another via texts or speech, etc.

Learning interventions can now harness the power and potential utility of the wider social

network. If we can spread ideas bi-directionally through these cultural networks in real time

then we can harness instant feedback and reuse. This will help create what can be termed

mobile meme machines (http://mememachine.com/). Application Program Interfaces (APIs)

exist for Twitter and Facebook which will make it easy to talk to Augmented Reality (AR)

browsers in real time. Students should be able to share and critique their ideas on urban or

school design both in situ and in reflection. Metadata could then potentially be gathered in

real time via these social networks and used to capture the meaning and sense-making

process in learning. All kinds of annotations, classifications, discussions, usage information,


and references can then be added as an extra layer on top of the content information. This

metadata can be used to enable users to find all kinds of new media for instruction and

learning (and potentially keep their Zone of Proximal Development on display). Metadata in

this sense is closely related and can be fed into the core processes of learning as reflection,

guidance, and feedback (Specht, 2009).

During the urban planning studies students learnt that the physical design of any educational

institution has a direct impact on the delivery of the education that takes place within that

context. Learning space design shapes our behaviour and influences our thinking. As a result,

a way of extending that understanding is to get the learner to participate in the design of that

context in order to control how their learning environment operates.

The idea of this application of the system is for urban designers working in pairs to use

simple open source 3D software (i.e Google sketch up http://sketchup.google.com/ or blender

http://www.blender.org/) to create basic additions or revisions to current school designs (in

the form of 3D sketches). This is a useful way to give students the opportunity to highlight

the areas of their institutions that in their opinion lead to negative effects on their learning.

In order to test the validity of the findings a virtual version of the application will be created

which will be used by individuals only. This will involve individuals using the recently

available Google Earth API (http://code.google.com/apis/earth/). In essence, instead of

placing their 3D models and annotations into real space they will enter them into virtual

space, via the AR application. This data will then be compared to examine how significant

the world as a platform is in this scenario.

Augmented Reality

Another way of extending the project is to build an Augmented Reality (AR) version of the

tour. AR specifically generates composite views using the real scene viewed by the user and a

virtual scene generated by the computer. The key aspect of AR is that the virtual elements

enhance the person's perception of the world by supplying relevant information that is not

contained in the real world.

AR is more powerful than VR as its influence is bi directional. “If we consider AR as a

visualisation technique, the relationship of real and virtual objects is one of focus and context:

Either we want to provide additional virtual context to an important object in the real world

or we want the user to focus on a virtual object embedded in a real context” (Kalkofen et al,

2009). The learner in an AR environment can locate points of interest (POI) which have been

pre-embedded into the scene or they can place (and eventually activate) their own POI into

the scene in real-time.

The construction of psychological and physical space is one of the constituent parts in the

generation of context. MAR (Mobile Augmented Reality) systems have been referred to as

“intelligence amplifying systems to enhance human cognitive activities, such as attention,

planning, and decision making” (Brooks, 1995).

AR crucially provides both the direct primary experience (the real world scene) and the

mediated representation (the digital augmentation). As a result Augmented Reality (AR)

provides significant support for real time situated learning. Practice and theory can now feed


off each other in the same space. A good example of this is an engineer working on an engine

whilst wearing glasses that provide a digital overlay showing the potential areas which could

be responsible for a known problem. The theory informs the practice in the same space.

Extending the Language Learning project

A lot of virtual media is often accused of isolating learners from reality but augmented-reality

has the opportunity to draw learners further into reality by amplifying the already existing

authentic context. The inherent high resolution nature of real life must be taken advantage of

in order to enhance learning. Language is fundamentally about the local context. As a result

the classroom is arguably not the most efficient environment for enhancing and assessing

language performance. Through acquiring a language in context you acquire a culture.

Authentic cultural language should directly mesh with authentic cultural situations.

As highlighted by Vygotsky’s Zone of Proximal Development (ZPD), maintaining successful

immersion in language learning depends on working at a level just above the comprehensible

level of the student. An average teacher has many students with differing levels of ability in

their language skills, and as a result, personalized immersion instruction is almost impossible

to apply.

Augmented Reality in this context provides an excellent vehicle for immersive and dynamic

language learning. Adding extra information onto the surrounding environment is the core

function of Augmented Reality. A language learner is supplied with the ability to acquire a

language whilst being augmented within their authentic environments. Learners can see local

people having conversations in the target language and real objects being labeled dynamically.

Preference settings could easily tailor the method to the individual level of ability ensuring

the learners remain within their ZPD.

Conclusion

The PLOTMaker is a versatile pedagogical tool where generated digital learning content can

be customised, adapted and edited for different learning purposes across a big range of topics

and disciplines.

Not only can teachers create content for students but also students can create content for other

students. This digital mentorship becomes a playful tool in the learning process. PLOTs can

be used alongside any other app running on the device to create effective ‘context

augmentation’ within a wide range of learner-centered experiences and possibilities. The

students can, for instance, create their own educational journey with the additional support of

Augmented Reality apps. Students can also create PLOTs dedicated to documenting the

process of creation of their projects.

Students can create engaging activities using their own original content based on the topics

they are interested in and instantly share them with their peers for iterative feedback. The

PLOTMaker authoring tool gives teachers and students a different approach to knowledge

construction. The pedagogical tasks can be chosen by the students, in an immersive


ubiquitous environment that can be discussed and enhanced through the use of social media

(i.e. Twitter or Facebook). References

Appadurai, A. (1996). Modernity at Large: Cultural dimensions of globalisation.

Minneapolis: University of Minnesota Press.

Brooks, F. (1995). The Mythical Man-Month: Essays on Software Engineering. Addison-

Wesley Professional.

Brown & Campione, 1994; Lave, 1991; Pea, 1993a; Scardamalia & Bereiter, 1996 in Stahl, G.

(2006). Group cognition : computer support for building collaborative knowledge.

Cambridge, Mass. : MIT Press.

Dawkins, R. (2006). The selfish gene. Oxford university press.

Durkheim, E. (1895). The Rules of Sociological Method. 8th edition, trans. Sarah A. Solovay

and John M. Mueller, ed. George E. G. Catlin (1938, 1964 edition) Kalkofen D., Mendez E., Schmalstieg D. (2007). Interactive Focus and Context Visualization

for Augmented Reality. In Proceedings International Symposium on Mixed and Augmented

Reality, Nara, Japan, pp. 191-200,

Kirsh, D. (1995). The intelligent use of space, Artificial Intelligence.

Kolb., D. A. & Fry, R. (1975). Toward an applied theory of experiential learning. In C.

Cooper (ed.), Theories of Group Process, London: John Wiley.

Pratt-Adams, S., Maguire, M. and Burn, E. (2010). Changing Urban Education. London,

Continuum.

Ryberg, T. & Christiansen, E. (2008). Community and social network sites as Technology

Enhanced Learning Environments. Technology, Pedagogy and Education. 17:3. 207-219.

Stahl, G. (2006). Group cognition : computer support for building collaborative knowledge.

Cambridge, Mass. : MIT Press.

Specht, M. (2009). Learning in a Technology Enhanced World: Context in Ubiquitous

Learning Support. Inaugural Address. September, 11, 2009, Heerlen, The Netherlands: Open

University of the Netherlands.

Sheller, M. and Urry, J. (2006) (Eds). Mobile Technologies of the City. London, Routledge.

17

Urry, J. (2007). Mobilities. Wiley.


Wenger, E. (2006). Communities of practice: a brief introduction.

http://www.ewenger.com/theory/

Won, J. Storkerson, P. (1996). Hypertext and the Art of Memory. Visible Language.

Yates, F. (1992). The Art of Memory. London. Pimlico.

Vygotsky, L. (1978). Mind in society. The development of higher psychological processes.

Cambridge, MA: Harvard University Press.


PLOTLearner as Persuasive Technology: Tool, Simulation and Virtual World for Language Learning

Nicolai Winther-Nielsen Fjellhaug International University College Denmark,

Frederiksborggade 1.B.1, DK-1360 Copenhagen K [email protected]

The Lifelong learning project EuroPLOT (www.eplot.eu) has developed PLOTLearner as

a new learning technology that uses a text database to drive persuasive language learning.

In terms of B. J. Fogg’s theory of Persuasive Technology from 2003, the function of

PLOTLearner can be characterized as a tool to use for the practice of language skills, but

also as a simulation of the interaction with the grammar of a text. It automatically

scaffolds the text with a virtual world consisting of pictures, videos, and documents.

PLOTLearner is furthermore developed according to Diana Laurillard’s model from 2012

of how to learn a practice capability from an external environment, which in this case is

a corpus for training and exploration.

A particular database of the Hebrew Bible from Amsterdam is used in this project, but

both the database management system and the PLOTLearner application are developed as

open source. The approach illustrated in this project could work for learning language and

culture from any text and for any language. This persuasive technology and its learning

objects is an example of a completely repurposable sustainability in learning technology.

Keywords: Persuasive Technology, Learning technology, Design for learning, Corpus-

driven language learning, Hebrew Bible database.

1 Introduction “Let the Hebrew Bible be your tutor!” This is the sales pitch used for a new learning

environment where a text drives the motivation and enablement of learners. This new brand of

learning technology, PLOTLearner, is one of two technologies developed by the lifelong

learning project EuroPLOT 2010-2013 (www.eplot.eu).

EuroPLOT has developed this technology for the small and manageable pilot case of learning

from the ancient corpus of the Hebrew Bible (Winther-Nielsen ms). However, the technology

could scale to any open project for learning to read from texts in any ancient or modern

language. The project has already experimented with texts in New Testament Greek, Danish texts

from the Kaj Munk collection and Frankfurter Rundschau texts from the German Tiger database.

Other papers from the project, and especially Behringer et al (2013), explain how EuroPLOT

has developed learning design and patterns for other areas of learning. This paper will

illustrate the principles of persuasive language learning that were developed for PLOTLearner

as a tool, medium, and virtual world for learning.

The acronym PLOT is short for Persuasive Learning Objects and Technologies. However, it

also hints at several senses of the verb ’to plot’, such as planning projects, mapping progress,

generating processes, and constructing plots in texts. These different usages capture the

structure of this discussion of PLOTLearner, explaining how

● the project planned a new kind of persuasive interface for an EU project (section 2)

● the tool maps a journey of learning through statistics and graphs (section 3)

● a learner plans self-directed learning from texts (section 4)


● a learning environment supports scaffolding for a better understanding of the plot of the

text (section 5)

2 Plotting a persuasive technology When EuroPLOT in 2010 proposed persuasive technology for language learning, we already

knew that quiz technology provided by the program 3ET can solve many basic needs of

beginners learning to read in a new language (Winther-Nielsen 2011). Understood in terms of

B.J. Fogg’s (2003) theory of Persuasive Technology, we already had a tool that could

‘motivate people to exercise’ (2003:32), but we lacked a stronger persuasive capability in the

technology to change the behaviour and motivation of users. We were persuaded by Fogg that

when a learning technology takes on the roles of pedagogical persuaders like teachers,

instructors or coaches, it may have great potential for education:

From the classroom to the workplace, educational designers will create computing

applications that deeply motivate people to initiate a learning process, to stay on

task, and then to review material as needed. ... As sophistication increases, we’ll

see adaptive education and training products that tailor motivational approaches to

match each individual learner – motivating ‘accommodators’ to learn through

cause-and-effect simulators, or providing ‘convergers’ with rewards for

performance on interactive problem sets and quizzes. (Fogg 2003:246).

Principles of persuasive technology are now finding their way into digital design. At the same

time edutainment and gamification are emerging as big industry in the marketplace, culture,

and entertainment. However, it is equally clear that sophisticated technology per se does not

guarantee effective learning or widespread use. The goal of our project was therefore to

explore how current work in design for learning, interaction design, and computer-assisted

language learning could inform our development of an application we called PLOTLearner,

with the aim of deployment and evaluation in a pilot project on a corpus of the Hebrew Bible.

Fogg (2003) characterizes the three core functions of a persuasive technology as tool,

simulation, and social actor. The following will explain how we are implementing those

functions in language learning, guiding learners to

● practice skills in reading, writing, and parsing for proficiency with a tool for practice

● explore the structure of a text for engagement in a simulation of the grammar of the text

● exploit resources activated by the text for a virtual world which functions as a learning

environment

In light of this, the aim of EuroPLOT is to develop a learning technology that empowers and

motivates self-directed learning driven by a database. Learners can use this technology to

explore the text through a text corpus interface. They can practice skills of reading and

vocabulary memorization and gain proficiency in morphology and in analysis of syntax and

text. PLOTlearner supports the construction, reuse, and repurposing of exercises which are

loaded, edited or created through a Windows interface. It generates activities for learning of

language or culture from a database storing a text corpus.

The programming of PLOTlearner has been structured as an agile development, with new

releases produced by Claus Tøndering for testing by Nicolai Winther-Nielsen about every

second to third month from April 2011 to December 2012. From April 2013 Claus Tøndering

is developing an online version without funding from EuroPLOT (http://pltest.3bmoodle.dk/).

The plan is to continue open source development after the end of the EuroPLOT project in


October 2013. Furthermore, we are now working on support for online display of the progress

of learners in a “Learning Journey Online” (Gottschalk and Winther-Nielsen ms).

The persuasion envisioned through PLOTLearner focuses on conditioning through

certificates, surveillance by teachers and peers, self-monitoring by self-directed learners, and

persuasive suggestions directed at autonomous language learners (Winther-Nielsen

submitted). The technology can be characterized in different ways due to its support for

language drilling, text display, reading, and vocabulary learning: PLOTLeaner is

● a tool to simplify the practice involved in acquisition of morphology (drills)

● a tutor that simulates the study of grammar in text (interactive display)

● an interlinear guide with transliteration (pronunciation support)

● a reading helper that checks typing, reading, and spelling skills (self-corrective reading)

● a translation assistant with glosses and ranking of word frequency (vocabulary training)

● a virtual world with display of cultural background material (interpretation)

The following section will explain how PLOTLearner uses the three functions of tool, tutor,

and virtual environment for practice and exploration into a corpus-driven learning scenario,

and how the interface is designed to enhance the persuasive effects during active learning of a

language and its texts and culture. The core of the present technology is explained from

Laurillard’s (2012) model of learning from an external environment.

3 Learning from a plotted practice environment

The step we first have to take is to plot how

the tool marks a course for skill training and

displays learner progress through statistics and

graphs. This tool aspect was already

developed for the technical quiz solutions in

3ET, but the goal now was to develop a

persuasive framework for practicing mastery of forms in a language. In Fogg’s terms, the

focus here was how to initiate, continue, and complete the practice of a skill by review.

The best way to explain the role of practice is to use the robust theory of design for learning

developed by Diane Laurillard. Her figure is reproduced with minor changes in Figure 1:

It shows the learner learning by using their personal goals and current conceptual

organization to select from their current practice to generate actions on the

external environment. The learner can use an action modeled by the teacher, or

use results from their own action to modulate and build their practice capability.

What they get from the teacher or the environment may modulate their concept,

their personal goals, or current practice capability, and so generate new actions in

continual iterative process. (Laurillard 2012)

Laurillard’s mapping of learning from an external environment explains how a persuasive

skill practice can emerge, and especially how, “the shading identifies the internal cognitive

components the learner deploys during the process, which the teacher is trying to influence”

(Laurillard 2012: 61). The figure clearly illustrates how the learners’ practice capability

depends on action modelled by the environment (BE). Crucially, in PLOTLearner the learners

interact with the external environment (E) through an interface to the database. However this


content is generated from concepts provided by the computational linguists who created the

database as well as by the course developers through conceptual organization (CT). In this

sense the learners’ action to achieve the desired practice skills (DE) is embedded within a

larger environment of facilitation. The persuasive effectivity is afforded by a simplification of

practice and by the enablement of a pleasant training environment that encourages learners to

improve performance to the level they want to achieve through their efforts.

(C ) Explanation T

(C ) Conceptual T

organization

(F ) Comment

T

(C ) L’s T

Conceptual

(D ) Articulation T

(A) L’s

Personal goals

(B ) Model action E

(E) External

environment

(F ) Result

E

(B ) L’s Practice L

capability

(D ) Action to E

achieve goal

KEY Generate Modulate

Figure 1: Learning from an environment and a teacher (Laurillard, Fig 4.1)1

PLOTLearner was developed and tested as a Windows program with two entry modes:

learners activate the program in the “PLOTLearner” mode and they are instantly brought to

exercises to train in a skill. Learning designers, teachers, and advanced learners can choose to

activate the program in its “PLOTLearner – Facilitator Mode”, which will allow them to edit

exercises or create new ones. They can provide descriptions of learning content or hints

addressing potential frequently-repeated mistakes and supply links to external content (see

http://www.ezer.dk/3ETusersguide/PL-2.0.1/en/intro.php). Each time the learner clicks on a

load exercise button, he or she is presented with a screen that enables him or her to study

information on the practice task in the "Exercise description" field, or else begin to practice

right away by pressing the “Run this exercise” button. The content of this description can be

1 Used by permission from Laurillard, including changes of her siglas: (CC), (FC) (DC), replaces (CT), (FT) (DT), because (CT) explanation can be confused with (CT) conceptual organization. Subscript letters ({C, D, F}C) are functions of the conceptual organization, hence by analogy ({C, D, F}E).


repurposed by any facilitator for any learner. In this way the facilitator can repurpose practice

to provide the exact amount of persuasive motivation to encourage engagement as well as

explain how learning routines can be simplified. Crucially, the description should define the

learning outcome and give easy access to whatever additional content with which the

facilitator wants the learner to engage.

For any language learning immediate feedback is crucial. PLOTLearner handles the Result

(FE) function by checking answers in the database and generates feedback. While the

facilitator does not have the ability to supply hints to address every possible wrong answer, he

or she can supply pedagogical advice on potential problems in the exercise description.

Learners receive instant feedback on their knowledge by directly checking answers or getting

feedback through the Result (FE) function. During the practice of a specific exercise for

language learning, the learner will always be able to press a “Check answer” button to

improve on the fly until the right answer is matched in the database. The learner can at all

times go back to the description and study some more, but this will have an impact on

performance speed. Another button will allow the learner to have an answer shown, but this

will count as a mistake. Otherwise, the learner can skip ahead to subsequent questions or end

the exercise. Throughout the process, a bar visualizes progress, showing unanswered

questions remaining.

Other sophisticated displays support the tracking of the learner’s journey, displaying the

duration and results of exercises and listing mistakes in a statistics files called “Complete

local report” (http://www.ezer.dk/3ETusersguide/PL-2.0.1/en/statistics.php). The learning

journey can be tracked in an “Exercise graph”, which graphically plots how many right

answers were given per minute. In this way there are several options available for display of

crucial feedback (FE). The value of these statistics and their use in tests and exams is

discussed in Winther-Nielsen (ms).

4 Self-directed learning plotted in a motivating interface

A further step on our path into this new

learning technology leads to the new

functionality offered in the second

prototype, PLOTLearner 2. Our route of

agile development has taken us far

beyond 3ET, which only supported

exercises on morphology and writing.

Earlier analyses of persuasive functions had indicated that we should support learning through

active exploration. Learners must begin with an overall understanding of the form of the

before they can be persuaded to learn actual forms. This enables the learner to gradually

acquire a higher level of mastery because the learner is in complete control of planning the

content, pace, and range of his or her learning journey through the interface. It clearly

indicates that PLOTLearner is not only a tool for practice, but also an adaptive technology

that implements Fogg’s category of persuasive media. It places the learner within a simulation

of the objects of the text, empowering the learner to gain full control of all active exploration.

In practice this implies that technology supports the learner’s interest in engaging

interactively with the text. This is the part where the learner will engage with PLOTLearner to

activate the conceptual learning process (CL). For this stage, the model action (BE) has to be

prepared through the teacher’s conceptual input for explanation of concepts (CT). In this


regard, PLOTLearner also offers what is available in more traditional teaching scenarios and

classrooms, However, it redefines the role of the teacher’s conceptual instruction as

supervision and coaching on how to use the content and structure of the database.

This implies that in PLOTLearner the learning of concepts is primarily structured as how to

learn the content of the main source of the corpus-driven learning and hence also how to use

the rich conceptual content stored in the database. Moreover, this part of the system is very

persuasive from a user experience point of view because the learner can point to language

features he or she wants to have visually highlighted. The relevant content from the database

is readily available as pop-up displays in a user-friendly mouse-over fashion. To study the

database content of a selected text segment is therefore very pleasant for the learner and the

interactive environment fulfills learners’ desire to have interactive information available at the

tip of their fingers. In terms of the analysis of the functionality of a Persuasive Technology,

this is the ultimate function of persuasive suggestion (Winther-Nielsen submitted, and briefly

summarized by Gottschalk and Winther-Nielsen ms).

However, the shift to a corpus-driven environment for the study of grammar has even more

profound implications. When we researched our new approach to language learning, we

discovered that persuasive language learning must start with engagement with the interface

into the grammatical categories stored in the database. PLOTLearner therefore radically

changes the curriculum and topics of language learning by enabling interactive access to the

text through the hierarchical levels of word, phrase, clause, and sentence and the features and

values stored at each level (http://www.ezer.dk/3ETusersguide/PL-

2.0.1/en/firstex.php?lang=he).

● At word level: ”spacing” is added to display individual lexical morphemes within words,

as well as transliteration to help beginners learn to read much faster. (Transliteration also

gives access for those who are not able to read the Hebrew script, but still want to learn

aspects of the grammar).

● Form in text lists the form of the root and affixes for easier display with the text.

● Lexeme offers access to dictionary words and vocabulary learning based on gloss(es) in

English, their frequency, and their part of speech.

● Morphology displays stem and tense of verbs, state of nouns, and person, number, gender

● At phrase level: phrase types in terms of part of speech and functions like predicate,

subject, and object are shown with labels.

● At clause level: clause type labels and text type are given.

Thanks to this new potential in corpus-driven self-directed learning, the next generation of

learners can now begin to learn lexical and grammatical categories through the interface.

They can get a first-hand interactive and visual feel for the hierarchal structure of the text, and

all language data supports the pursuit of personal interests and goals. Evaluation data supports

our assumption that engagement with the text through an interface motivates learners to

initiate skill training for acquisition of forms, vocabulary, and syntax.


5 Corpus-solutions enabling language learning plots

The last step on our tour is to

examine how this new kind

of database-driven persuasive

technology can support even

more effective and efficient

learning in and around a

database by triggering the

engagement of learners. Fogg

(2003) has pointed to the

crucial role of khairos, or the

opportune moment for

learning, and the challenged

is how this kind of timing

can be triggered for

persuasive language learning

from a database.

Current language learning theory has moved away from communicative language teaching

and is heading towards task-based language teaching. The focus in this approach is on learner

defined tasks that support the understanding of the form of the text rather than memorization

of rare forms (Robinson 2011). The theory of Task-based Language Teaching has for long

focused only on development of a curriculum for the classroom and not paid attention to

learning technology. However, PLOTLearner joins a new trend of applying computer-assisted

language learning to task-based language teaching, such as seen in the studies published by

Thomas and Reinders (2010). The ultimate goal is of course to let the learning technology

generate scaffolding that activate content for the Vygotskyan zone of proximal development

PLOTLearner does not support this ultimate goal yet, but we have made a start by creating a

flexible database system that can store pictures, videos, and documents for use in problem-

based learning. Whenever PLOTLearner is opened, the program will automatically connect to

the internet and activate the database EuroPLOT Resources

(http://resources.3bmoodle.dk/img.php). This database is developed as a kind of open source

learning repository that allows learning designers to construct their own content for the text

collections they want to include. In the EuroPLOT project, PLOTLearner is set up to

automatically search in the metadata for any resource that is linked to a text or to a

geographical reference mentioned in this particular text, but the pictures additionally contain

information on archaeological periods and cultural customs. A developer of learning content

can create any kind of new sub-topics and add any kind of content for this kind of intelligent

scaffolding of the text by multimedia. The filenames refer to the name of the creator of the

content, the category of the learning material and resolution types. It creates a visual world

that triggers a pedagogical visualization for persuasive learning and can be used for problem-

based learning activities. PLOTLearner automatically scaffolds this for the texts and displays

the metadata that are provided for learning objects.

For the present test of the learning environment, the bulk of the content in the database is

some 5000 pictures with metadata primarily in Danish. However, this database will be used

for ongoing experimentation with persuasive learning objects offered as open educational

resources for scaffolding with PLOTLearner and content provided in English.


Ultimately we want to go beyond suggestions generated by the text database and move

towards an intelligent tutoring system with tailored triggers. Our ultimate dream is to

construct a learning environment that scaffolds a deeper understanding of the world of the

text. Through further open source development of PLOTLearner, we aspire to support reading

for the plot from the beginning of a text through its unfolding in dramatic turns and peaks and

all the way to the resolution of the problem and the end of the story.

Conclusion This concludes the tour of the learning environment of PLOTLearner, with the introduction to

an entirely new way of learning a language from a text database. The journey included the

story of the birth of the project as well as a description of the crucial hallmarks of

PLOTLearner as a Persuasive Technology. In terms of the theory of Fogg (2003), it started as

a tool for practice of language skills and was developed into a simulation for the study of

language in texts. The social world is so far only included as part of pedagogical texts,

images, and videos automatically displayed by the program.

This sophisticated learning technology is developed as a highly adaptable and repurposable

open educational software. PLOTLearner is constructed to allow for inclusion of any kind of

text in any kind of language that has been stored in an Emdros database. It is our firm

conviction that this novel kind of persuasive technology has great potential for future

persuasive corpus-driven learning.

Acknowledgement The EuroPLOT project was funded by the Education, Audiovisual and Culture Executive

Agency (EACEA) of the European Commission through the Livelong Learning Program with

grant #511633. I would like to thank programmer Claus Tøndering for indispensable and

unfunded programming since July 2012 and Judith Gottschalk for challenging comments.

Language editing by Randall Tan has been invaluable.

References

Behringer, R., Gram-Hansen, S..B., Soosay, M., Mikulecka, J., Smith, C., Winther-Nielsen, N., & Herber, E..

(2013). Persuasive Technology for Learning in Business Context. International Journal of Information

Systems and Engineering, 1(1).

Fogg, B.J. (2003). Persuasive Technology. Using Computers to Change What We Think and Do. San Francisco:

Morgan Kaufmann.

Gottschalk, J., & Winther-Nielsen, N. (ms). Persuasive skill development: On computational surveillance

strategies for modeling learning statistics with PLOTLearner. Submitted to IWEPLET13.

Laurillard, D. (2012). Teaching as a Design Science: Building Pedagogical Patterns for Learning and

Technology, p. 60. Routledge, London.

Robinson, P. (2011). Task-Based language learning: A review of issues. Language Learning, 6, 1-36.

Thomas, T, & Reinders, H. (2010). Task-based language learning and teaching with technology. London and

New York: Continuum.

Winther-Nielsen, N. (2011). Persuasive Hebrew Exercises: The Wit of Technology-Enhanced Language

Learning. In W. T. van Peursen & J. W. Dyk (Eds.), Tradition and Innovation in Biblical Interpretation.

Studies Presented to Professor Eep Talstra on the Occasion of his Sixty-Fifth Birthday (pp. 277-298).

Leiden: Brill.

Winther-Nielsen, N. (submitted). The Corpus as Tutor: Data-driven Persuasive Language Learning. Submitted to

Digital Humanities Quarterly.

Winther-Nielsen, N. (ms). PLOTLearner as Persuasive Technology: Tool, Simulation and Virtual World for

Language Learning. Submitted to IWEPLET2013.


Chemical Handling – PLOTMaker for Training in Exposure Scenarios under REACH

Margrethe Winther-Nielsen and Ilaria De Rosa Carstensen

DHI, Agern Allé 5, DK-2970, Hoersholm, Denmark

[email protected], [email protected]

Persuasive design of e-learning is a relatively new area and not widely applied in training

directed at industry and business. The Lifelong Learning project EuroPLOT has

developed an e-learning resource on how to handle exposure scenarios for chemical

mixtures according to the EU chemical legislation REACH. Eight sessions were

developed applying persuasive principles. Each session consists of several persuasive

learning objects. The course was uploaded on a Moodle website and offered to the

industry as a stand-alone course. Sessions of the course were pre-tested in a focus group

at DHI and a pilot-test was initiated in July 2013. Employees from industries, branch

organisations and consultants were invited to the pilot-testing. The conclusion based on

the first pilot-testing results was that training in a challenging topic such as the practical

implementation of new chemical regulation by industries may be better understood

through simplification by reduction, tunnelling, tailoring and application of animated

interactive exercises. The possibility for learners to explore a topic at their own pace was

found motivating. It was, however, obvious that some learners may benefit from a

blended learning course or an online learning environment supporting discussions and

contact with a teacher.

Keywords: Persuasive design of e-learning, formulating industry, exposure scenarios,

case study

1 Introduction In 2007 the European Union radically changed its chemical regulation with the introduction of

REACH. The responsibility for management of the risk of substances was now placed on the

industry, and this new strategy for the EU chemical policy introduced entirely new obligations

for the industry. As a result, manufactures and importers must register and make a chemical

safety assessment, depending on the tonnage of a chemical substance they are producing or

importing. As part of the chemical safety assessment they may develop the so called exposure

scenarios, which we are dealing with in our case.

Different roles are assigned to the industrial companies and the professional users of

chemicals. The most important role for our case is the formulator, who is the downstream user

of chemicals producing chemical mixtures, such as paints, cleaning agents and glues. The

formulators must deliver information on the safe use of their products to their customers,

applying the information in the exposure scenarios annexed to safety data sheets received

from their suppliers of chemicals. Not only is it a challenge for the formulators to work with

the huge amount of information embedded in exposure scenarios, but scenarios are often

delivered in formats varying from supplier to supplier.

The main objective of our work was to create persuasive learning objects that addressed the

health and safety specialists within the formulating industries or consultants working with

these industries. Persuasion is of course not a new competence in teaching, but there have

only been few attempts to combine persuasion with e-learning for education of the staff in

business and industry applying the basic principles of persuasive technology defined by Fogg

(1998 and 2003).


Fogg argued that a persuasive technology may function as a tool for making behavior easier, a

medium for simulation or as a social actor. For each of these roles he designated a list of

persuasive principles (Fogg 2003; Gram-Hansen 2012). While the principles of Fogg

primarily are directed to persuasive technology, Oinas-Kukkonen and Harjumaa (2009)

extended these principles to a description of 28 different persuasion technics in their model of

Persuasive System Design (PSD). Oinas-Kukkonen (2010) combined this model with analysis

of the context of persuasion focusing on the designer and the usability of the user.

Persuasion is defined as the attempt to change attitudes or behaviours (Fogg 2003). By using

persuasive design in an e-learning resource on exposure scenarios, related to the chemical

legislation REACH, we attempted to engage and motivate the learner to actively apply the

information on safe use of chemicals included in exposure scenarios delivered to him from his

suppliers.

In this paper we present our work on developing the e-learning course, starting with course

planning, continuing with design of learning sequences and persuasive learning objects and

concluding with the first results of our pilot testing.

2 Materials and methods applied in the case study Tools for developing persuasive learning objects

PLOTMaker was the main tool applied in the development of our learning objects. The tool was developed within the EuroPLOT project and based on the software tool GLOMaker

developed by the Learning Technology Research Institute at London Metropolitan University

(http://www.glomaker.org/index.html). Both the Freestyle and the EASA pedagogical patterns

of the planner part of the software were used in the creation of our learning objects.

The PLOTMaker provides a teacher oriented authoring tool for creating persuasive learning

objects. By incorporating new layouts and interactive components it has become a suitable

tool for designing more engaging e-learning resources. Some of the components of the tool

are in themselves interactive while other components allow importing of interactive flash

animations and sounds created outside PLOTMaker. These resources are integrated into the

Learning Object (LO) in its final shape.

We created several interactive learning objects in our sessions applying Microsoft PowerPoint

in combination with iSpring Suite or Camtasia Studio. All sessions were assembled as a

guided learning tour in how to handle exposure scenarios for chemical mixtures. The entire

course was delivered in a Moodle website. Persuasiveness in the design

We started designing learning objects applying the persuasive principles defined by Fogg (2003) and exploring how PLOTMaker could support interactive e-learning. Our primary goal

was to design LOs, which motivated the learner to continue learning all the way to the end of

the e-learning course. Furthermore, the learning content should inspire to further studying of

the subject. It should provide employees from the formulating industry with a more

comprehensive way to learn the basic principles of working with exposure scenarios for

mixtures. The persuasive principles applied were: tunnelling, tailoring, reduction, simulation,

self-assessment, suggestion, and to some extent social signals. Testing by learners

A focus group at DHI tested selected sessions of the course in a pre-pilot test. Their

experience of learning with the LOs was gathered in interviews directed by Herber (2012).

Invitations to take part in pilot-testing of the whole course were sent to formulating industries,


branch organisations and consultants in Europe. When completed the online course each

participant was asked to file in an online evaluation questionnaire anonymously. The

questionnaire was developed by Erich Herber in 2013.

3 Results and discussion Planning the course

Selection of target group, overall topic and learning outcome are of course essential parts of the planning of a course. These tasks should all be defined prior to the detailed planning and

design of the course. For our specific case, a need expressed by industries was our basic for

the selection of the particular topic and target group. The typical target of the learning is often

one or two employees in each formulating company. We suggested that it would be beneficial

to them to have access to an online course covering the whole topic.

A team of three teachers at DHI selected the course content and took part in the design of the

course. The selection of content was based on our existing experience gained in several face-

to-face courses and workshops within the area. Additionally, two other teachers were

involved in evaluation and quality assurance of content and design.

The steps for the planning and design of the course were:

1. Selection of other software tools besides PLOTMaker which was preselected for our case

in EuroPLOT and a suitable learning management system (LMS)

2. Selection of course content, number of sessions and topic for each session

3. Learning sequence, including pedagogical pattern within each session

4. Design and selection of materials for each LO

5. Internal evaluation and quality assurance of content and functionality of LOs

As already mentioned we chose to apply tools to supplement PLOTMaker in order to obtain a

more varied and engaging learning. Knowing safety data sheets and classification rules for

chemicals substances and mixtures were set up as a prerequisite for taking the course.

However, the new terminology and methods introduced in the chemical legislation in 2007

were not assumed to be known in advance.

Before starting the design of LOs we decided on the sequences of learning activities for a

session, applying the planner in PLOTMaker. We found both the EASA and Freestyle

patterns in the new version of the software tool well suited for our purposes.

In total eight separate sessions were created, each consisting of a range of LOs. The content of

the LOs was mostly developed from scratch and only few objects were reused or repurposed

from LOs of others. Each session was packaged and uploaded on the Moodle website with a

link to the GloWebPlayer.swf file. The uploaded course was discussed and evaluated

systematically by the team and at least by one teacher outside the core team, and the course

was adjusted before release for testing.


Persuasive learning objects

The front page of the course was inspired by the Learning Journeys (http://hermes.uwl.ac.uk/learnerjourney/), guiding the learner through the different sessions of

the course (Figure 1), and providing an overview of the entire course. A learner without any

prior knowledge of the subject may follow the suggested route from session one to eight

while more experienced learners may jump to the sessions needed and thus plan a tailored

learning tour.

Figure 1: The front page of the online course The topic of exposure scenarios is in general considered quite complex learning content. The

learning has so far been achieved through reading of technical guidances and face-to-face

workshops and seminars. To our knowledge this is the first attempt to make an online course

on this particular topic. When transferring the teaching to the PLOTMaker software we were

able to reduce the big amount of guidance text to be read by presenting new terms and core

knowledge in simplified illustrations and animations. One example is the LO in Session 2 of

the course shown in Figure 2. By persuading learners to press the play button in this LO they

will be presented with text and pictures related to a specific term or descriptor, e.g. the SU9.

However, by linking to the original guidance in other LOs, the learners will still be able to

study this in details.

According to Fogg (2003) we are likely to be persuaded by something which appears to be

easy, and one important step towards making something easy is to make sure that everything

seems familiar to the users (Gram-Hansen 2012). Selection of situations, illustrations,

language, pictograms and other material that are known to the target group is a simple way to

make a LO familiar to the learner.

The illustration at the front page of the learning resource (Figure 1) appeals to the target

group, illustrating entities from their world. When entering the sessions you meet illustrations

of situations where people are using chemical mixtures or products, such as professionals

spray painting a car and technicians handling chemicals in a laboratory. Interviews of

colleagues from other formulating companies who express their opinion on a topic is another

way of making the LO familiar and the topic relevant to the learner. Besides including

opinions of colleagues in LOs, social signals were introduced by using pictures of people in

familiar situations.


Figure 2: Example on simplified presentation of new terms in an animation Self-assessment and simulation take up large part of the learning sessions, allowing the

learner to learn through interaction. The Multiply Choice Questions, Matching Quiz and Drag

and Drop components of PLOTMaker have been used several times throughout the course to

encourage learners to test their knowledge. Animations have been produced without

programming, creating sequences of pictures in PowerPoint which were converted to flash

files using the iSpring Suite. Use of animation in quizzes allowed inclusion of unexpected

effects, like a simulated explosion when the learner chooses the wrong answer (Figure 3).

Figure 3: Animation used in a quiz to make a special effect when the answer is wrong Simulation in our case means working with examples of exposure scenarios for chemical

substances, going through exactly the same steps that health and safety specialists in the

formulating industry have to perform when they receive real exposure scenarios from a

chemical supplier. Before learners enter into a situation of simulation, they are offered the

option to go through an example presented in a video. The simulations in the last part of the

course (Sessions 7 and 8) allow learners to do a calculation using an interactive calculator

(http://glomaker.wikifoundry.com/page/Links+to+free+resources?t=anon) and to type results

directly into the LO. Having completed the steps needed for preparing an exposure scenario

for a chemical mixture, learners in the end obtain the exposure scenario of the mixture.


Testing and interview of focus group

The focus group participating in the pre-testing was asked to think back on how they learned with the digital learning resources and how they benefited from them. The learners

commented on individual LOs and gave suggestions for improvement. They used the

feedback given in LOs repeatedly during the test to reflect on their individual learning

success. However, the feedback was not in itself considered motivating for the learning.

The application of video-based interactive exercises was considered useful by learners and

helped them to understand the exercise. Keeping simplicity in the LOs was recommended,

although exercises should remain challenging and demanding. In some LOs, text and image

were not synchronized and this caused irritation. The learner had to press a button for

changing the text in the right part of the screen as the animation went on to the left.

Unfortunately, automated synchronisation was not supported by PLOTMaker. The general

lesson learned from this is perhaps that one should not try to push a system or tool beyond its

capability, but keep it simple and functional.

The pilot testing of the course was still ongoing while this paper was prepared (August 2013).

Examples of the first preliminary results from six questionnaires indicated that the majority

(83%) enjoyed training with the online course, that the course helped them to better

understanding the practical relevance of the subject (100% i.e. strongly agree / rather agree),

that training with the online course can be useful for learning the subject (83%), that they

could find effective ways to learn with the online course (83%), and that the online course

motivated them to study the subject in more depth (67%). Furthermore, 67% of the testers

disagreed on the question “I believe that I can better learn from teacher instruction than from

online course”.

Four out of the six testers found that the online course was more motivating than other forms

of teaching and training, as expressed by their answers and comments (Table 1). They

appreciated the possibility to explore the topic at their own pace and without interference

from other people. However, a drawback of online learning could be the lack of feedback and

possibility to ask questions.

One of the testers pointed out that he or she experienced some technical problems during

testing, because the connection slowed down during browsing in the online course. It was also

annoying that a session started from the very beginning when he or she attempted to refresh

the connection. Furthermore, comments on some of the soundtracks pointed out that it could

be difficult to hear the content.

The pilot testing is anticipated to include at least 15 vocational learners from the industry. But

we are still waiting for the remaining replies. However, from the answers presented above it is

indisputable that learning through the online course motivated the majority more than face-to-

face teaching. The first step in fulfilling our overall goal of the online course - “to engage and

motivate the learner to actively apply the information on safe use for chemicals included in

exposure scenarios…” - is achieved because testers so far has found that the course has helped

them to a better understanding of the practical relevance of the subject.


Table 1: Answers and comments from six testers in the pilot testing of the online course

Question to the testers:

Was training with the online course more motivating than other forms of teaching & training (e.g.

classroom teaching / face to face training)?

Answer Number of

testers

Explanation / comment of testers

Absolutely 1 No disturbance and not 100 people asking questions about all and nothing.

Rather yes 3 1. Can keep my own pace, can jump back and forth as needed, can test myself w/o risk of being embarrassed.

2. Own pace

3. You can choose when to go in depth with a question

and when to go lightly on a subject.

Rather no 0

Not at all 1 No feedback, no possibility to ask questions

Indecisive 1 The training I received earlier was rather poor (in a workshop) this did not inspire me to attend more courses.

But courses/classroom teaching is ok, but if things are very

complicated, much time is spent on the poorest participants.

The way I learn is to have some background (like this

program or to read the theory), then ask questions and listen

(face to face training or classroom training).

4 Conclusions

This paper demonstrates that the principles of persuasive design can be applied in e-learning

directed at industry and business. We have shown that training in a challenging topic, such as

the practical implementation of the new chemical legislation by industries, may be better

understood through simplification by reduction, tunnelling and tailoring of the subject.

The application of video-based or animated interactive exercises support learners and keep

them engaged in their learning. The possibility for learners to explore a topic at their own

pace was found motivating. However, it was obvious that some learners may benefit from a

blended learning course or an online learning environment supporting discussions and contact

with a teacher.

Acknowledgements The EuroPLOT project was funded by the Education, Audiovisual and Culture Executive


grant #511633. We would like to thank partners, colleagues and clients in the industry for

their contributions.

References

EuroPLOT (2010). Persuasive Learning Objects and Technologies: http://www.eplot.eu/

Fogg, B.J. (1998). Persuasive Computers, Perspectives and research directions. in CHI. 1998: ACM Press

Fogg, B.J. (2003). Persuasive Technology – Using Computers to Change What We Think and Do. San

Francisco: Morgan Kaufmann Publishers.


GLOMaker. Links to free resources: http://glomaker.wikifoundry.com/page/Links+to+free+resources?t=anon

(2013)

Gram-Hansen S.B. (2012). Persuasive Learning Design Framework. Persuasive Learning Designs Report for the

EuroPLOT project. Deliverable D3.3.

https://sites.google.com/site/llpeuroplot/D.3.3%20PersuasiveLearningDesigns.pdf?attredirects=0&d=1

Herber, E. (2012). EuroPLOT Focus Group Meeting. Internal document of EuroPLOT.

Learning Journeys: http://hermes.uwl.ac.uk/learnerjourney/ (2013)

Oinas-Kukkonen, H. & Harjumaa, M. (2009). “Persuasive Systems Design: Key Issues, Process Mod-el, and

System Features” in Communications of the Association for Information Systems, Vol. 24, 2009, 28, p. 485-

500. The Berkeley Electronic Press.

Oinas-Kukkonen, H. (2010). Behavior Change Support Systems: A Research Model and Agenda. In Thomas

Ploug et al, eds, Persuasive 2010. LNCS 6137. Berlin Heidelberg: Springer-Verlag, 4–14.


Applying persuasive principles to influence students in adopting deeper learning approaches

Mekala Soosay

Leeds Metropolitan University, Leeds LS6 3QS, United Kingdom

[email protected]

Jaroslava Mikulecka University of Hradec Králové, Hradec Králové, Rokitanskeho 62

[email protected]

The current generation of students are exposed to an increased use and familiarity with

communication, media and digital technologies. In line with this capability, universities

are offering an arena for engagement and discovery, making use of the Virtual Learning

Environment (VLE) to provide a blend of online and face-to-face curriculum. Persuasive

design is a relatively new framework for influencing behaviour, made prominent by BJ

Fogg (2003) where he proposes a set of persuasive principles on how to think about creating

designs that are more persuasive. This paper articulates the development, evaluation and

sharing of PLOTs in teaching database concepts to students at Leeds Metropolitan

University (Leeds Met) and the University of Hradec Králové (UHK). Students of

undergraduate Computing and Business Computing studies face challenges in

understanding fundamental database concepts. In order to develop competency, we have

applied selected persuasive principles in the design of blended learning using persuasive

learning objects (PLOTs), self-assessment and discussion boards within the institutional

VLE. Students participated in the evaluation of the PLOTs, engaging in discussions with

peers and tutor, and self-assessing their understanding of concepts. Results indicate a

degree of behaviour change towards becoming more self-regulated learners.

Keywords: persuasive design principles, persuasive learning objects, self-regulation,

blended learning.

Introduction This paper discusses the development, evaluation and sharing of Persuasive Learning Objects

(PLOTs) in teaching database concepts for Computing and Business Computing students at

Leeds Metropolitan University (Leeds Met) and the University of Hradec Králové (UHK). The

initiative forms part of the EuroPLOT project, contributing to the understanding of the cross

field between persuasion and didactics. Persuasion is not a new concept in education; tutors

have used it in various forms to influence students’ learning behaviour in the classroom, and

the challenge is to apply the concepts effectively through digital e-learning technologies. BJ

Fogg (2003) paves the way in providing the structure around which people understand and

apply persuasive design, mediated through computer technology.

A digital learning object is defined as an online interactive chunk of e-learning designed to

achieve a stand-alone learning outcome, which can be reused in different contexts. When

persuasive design principles are applied to the creation and deployment of the learning objects,

they are referred to as PLOTs (persuasive learning objects) with the main objective of achieving

active engagement. At both universities, an outcome-based learning design paradigm is used

for creating the courses, hosted on the Blackboard Virtual Learning Environment (VLE) and

delivered as a blended learning context. Garrison and Kanuka (2004) posit that when blended

learning is thoughtfully integrated into face-to-face (F2F) learning, the strengths of


synchronous (F2F) and asynchronous (text-based Internet) learning activities can be harnessed

to support meaningful learning outcomes. Gram-Hansen (2010) points out that there are shared

commonalities between key persuasive design concepts and outcome-based learning design

(Table 1), which we have incorporated in the development of our PLOTs.

Table 1. Mapping of key Persuasive Design concepts to Outcome-based Education (OBE)

learning design

Persuasive Design OBE Learning Design (Biggs & Tang, 2007)

Originates from a persuasive intention

Considers the requisites of the users

Requires that the users is aware of the

persuasive intention

The persuasive intention is met through use of

one of more persuasive strategies

Is dependent on timing and contextual

awareness

Originates from an intended learning outcome

Considers the requisite of the students

Requires that the students are aware of the

intended outcome of individual lectures and

courses

The intended learning outcome is achieved by

use of rhetorical and didactic strategies.

Is dependent on timing and contextual

awareness.

The application of such principles raises the concern whether these can enhance the learning

experience of students accessing and using the learning objects within the VLE, specifically

instigating a change in attitude, understanding, or behaviour (Fogg, 2003). Within Higher

Education (HE) curricula, this translates as an opportunity to promote self-regulation through

independence and personal ownership of learning, increasing students’ ability to self-assess and

self-correct, skills which form the hallmarks of undergraduate education. These elements also

point towards self-efficacy, a quality which the prominent psychologist Albert Bandura (1997)

refers to as “beliefs in one’s capabilities to organise and execute the courses of action required

to produce given attainments” affecting students' motivation and learning, and self-efficacy

forms an important aspect of contemporary educational psychology (van Dinther, Dochy &

Segers, 2011).

Teaching of databases is integral to all Computer Science and Information Technology-related

courses. Graduates should be self-efficacious in solving real-world problems, which require

combining the knowledge of fundamental database concepts with the ability to recognise when

a particular problem could benefit from the application of specific concepts. This process

enables students to gain database skills that are valued in the workplace. The Teaching,

Learning and Assessment in Databases (TLAD) which is an UK annual event that brings

together database teachers and researchers in all academic institutions globally to share good

learning, teaching and assessment practice, acknowledges that teaching concepts such as

normalisation and Structured Query Language (SQL) are problematic. Hence we have chosen

to develop PLOTs supporting the understanding of normalisation at Leeds Met and SQL at

UHK.


Pedagogic and Persuasive Underpinnings in a Blended Context Knight’s longitudinal study (2010) which evaluates the different learning strategies adopted by

students when accessing VLE-hosted resources, reveals students who adopt a deep learning

approach, in which online resources are accessed consistently throughout the study programme,

perform markedly better than surface learners who focus their online activity either at the

beginning or end of the programme’s duration. When analysing the behaviour of our students

in the VLE, it was apparent that they adopt a ‘grab-and-go’ strategy. The challenge was to

enhance the course with PLOTs, designed to afford students to engage with learning actively

throughout the semester. We believe that in order for the PLOTs to suitably change students’

behaviour towards adopting a deeper learning approach, an appropriate activity blend ought to

be designed embedding persuasive principles, besides incorporating these principles into the

design of the PLOTs themselves.

Persuasive Design Principles viable in VLEs Fogg (1999) proposes a framework for persuasion, which he refers to as the Functional Triad

(Table 2), in which the computer as a tool can increase capabilities by enabling persuasive

affordances that develop self-efficacy. This is in line with Bandura's (1997) pioneering theory

where self-efficacy is a perceived quality; that even if individuals sense that their actions are

more effective and productive (through using a specific computing technology), they are more

likely to perform a particular behaviour, suggested as persuasive affordances below (Fogg,

Cuellar & Danielson, 2003). We will adopt the Functional Triad as a starting point in designing

our PLOTs.

Table 2. Specific Persuasive Affordances arising from the use of computer technology,

adapted from Fogg (1999)

Function Specific Media

within the VLE

Specific Persuasive Affordances

Computer as tool

or instrument

PLOT

Reduces barriers (time, effort, cost)

Increases self-efficacy

Provides information for better decision making

Changes mental models

Computer as

medium

PLOT

Provides first-hand learning, insight,

visualisation, resolve

Promotes understanding of cause-and-effect

relationships

Motivates through experience, sensation

Computer as

social actor

Discussion Boards

Establish social norms

Invoke social rules and dynamics

Provide social support or sanction

The intended change of students’ behaviour is classified according to the Fogg and Hreha’s

(2010) behaviour wizard as BlueSpan, whereby we expect students to do a familiar behaviour


over a period of time. The behaviour wizard suggests that designing span behaviours requires

special consideration where a span intervention might pay close attention to the strategic use of

regular triggers which have to be activated when the person has sufficient motivation and ability

to perform the behaviour. The VLE allows for announcements and notifications that can be

used to prompt students to undertake learning activities at appropriate points. As the VLE keeps

track of the individual student’s access of activities, timely triggers can be issued when most

students are online to get them to engage in the learning activities. The activity blend, which

includes the PLOTs is designed with the aim of allowing students a chance to quickly review a

topic, engage in collaborative discussions with the tutor and peers, and embark on formative

assessment designed to improve self-efficacy. Sadler (1998) refers to formative assessment as

assessment that is specifically intended to generate feedback on performance to improve and

accelerate learning. Nicol and Macfarlane-Dick (2006) take this further in revealing the

positive implications of using formative assessment for students to proactively generate and use

feedback in supporting the development of self-regulation.

The four persuasive principles (Fogg, 2003) adapted and applied to the design of our PLOTs

and VLE activity blend are:

Reduction - simplifying a complex concept through PLOTs.

Suggestion - intervening with a trigger, as compelling suggestions at the opportune moment

through the discussion board and emails.

Self-monitoring - regulate progress of engagement and learning with PLOTs through the

self-assessed quizzes.

Social signals - rewarding users with positive feedback and providing support through the

self-assessment and the discussion board.

We have applied a pedagogical approach of scaffolding learning to support achievement of the

learning outcomes as advocated by McLoughlin and Lee (2010), provoked by triggers at

appropriate points of learning guide students in a step-by-step manner, allowing them to

independently construct their learning in a way most suited to the individual needs. The PLOT

content is structured to be gradually absorbed by the learner using scaffolded delivery of

concepts in increasing complexity (McLoughlin and Lee, 2010). We used different forms of

media for motivating learner engagement in terms of capturing and focusing attention such as

audio, images and video. Researchers and cognitive psychologists believe that technology can

deliver more sensory data than we can process, to the extent that audio and visual effects can

interfere with learning (Clark and Mayer, 2011). Care is taken to avoid the redundancy effect,

where animation is combined with supporting narrative, but not with corresponding text that

can hinder associating words with corresponding graphics. On the same note, extraneous use

of text and graphical images are also avoided, which are in line with persuasive principles.

This is followed by short formative assessments in the form of quizzes for students to self-

assess, so that they can ascertain if indeed learning has taken place. The premise is that students

self-monitor their learning, capitalising on the flexibility of the medium that allows for iterative

any time, any place exploration until they have understood the database concepts delivered

through the PLOTs by self-assessing their knowledge gained.

The PLOTs are hosted on the institutional Blackboard VLE to facilitate a constructive blend of

learning designed to inculcate self-regulation in students. The VLE also affords asynchronous

discussions as threads that are necessary in depicting interaction and presence. Garrison and

Kanuka (2004) find that these elements are alluded to by students as important aspects of online

course design especially in instilling a sense of community. The proposed activity blend

involving a persuasive pedagogical model for delivering learning via PLOTs are as follows:


1. Understand the guidelines and procedures for using the activity blend,

2. Work through the PLOTs,

3. Discuss tasks and issues in the discussion board with tutors and peers, and

4. Undertake the formative assessments.

Case Study Description The PLOTs are designed to be used independent of tutor guidance, supporting the F2F teaching

and learning of the relational database design and the database query language SQL in 2

modules:

Introduction to Database (IDB) - first year undergraduate module offered to BSc.

Computing and BSc. Computer Forensics students at Leeds Met, and

Database Systems 1- second year undergraduate module offered to Information Management

and Applied Informatics students at UHK.

The main aim of PLOTs are to persuade students to revise or recall topics as the bite-sized

stand-alone learning objects, whereby the main concepts are embedded into an appropriate

curriculum and taught F2F. Topics correspond closely with the course learning modules, and

each topic has clearly defined learning outcomes, which are assessed within the PLOT and the

VLE.

Evaluation Findings Findings arising from the 3 focus group questions posed to students at Leeds Met are

summarised as follows:

1. What is your favourite / most useful learning feature or exercise in the learning object?

Students remarked that the videos afforded interactivity that enabled understanding of complex

normalisation concepts, allowing them to control the effectiveness of learning and reinforce

understanding deeper, and the much-needed rehearsal for learning (Clark & Mayer, 2011).

They felt that the videos condensed F2F teaching materials into concise chunks of learning, and

allowed for more inclusive learning – both online and during F2F lab sessions.

Students pointed out that the audio narrative clarified and simplified understanding of concepts,

and the voice humanised the PLOT, motivating them to engage with the content. The audio also

strengthened trustworthiness and credibility (Fogg & Tseng, 1999) which supported the

portrayal of a good teaching presence online (Garrison, Anderson & Archer, 2000). Students

found the quizzes within the PLOT useful as it allowed them to self-assess the extent of

understanding the database concepts

2. Which problems (technical/didactical/organisational) did you have? What caused them?

Students felt that the PLOT was easy to navigate and use. They did not experience any

pedagogical issues in terms of understanding the concepts explained in the PLOT, which is designed to expose concepts in a progressive and scaffolded manner. Students did face

technical problems in accessing the learning object from the VLE which was eventually

resolved by downloading the Flash plug-in.

3. If you could add any new feature or exercise to the learning objects, what would it be?


Students highlighted the need for more interactivity in the assessment, whereby they could

complete exercises that will drag and drop attributes to build normalised tables interactively.

They also expressed the desire to have clues/tips to help when incorrect choices are made under

the quiz section in the PLOT.

Similar findings were noted with UHK students using the same three focus group questions.

Question 1 - Most students agreed that the favourite/most useful learning feature in the learning

objects were the quizzes, which they have used as many times they wished, thus helping them to increase self-efficacy (Bandura, 1997). They found that the practical examples and the

scaffolded delivery of learning in small chunks clarified their understanding of concepts.

Question 2 - Students experienced technical issues in displaying the study materials and quizzes

in low resolution when viewed with a laptop, as well as issues with JavaScript that prevented

the VLE from functioning properly. Consistent with the findings from Leeds Met, students

suggested that the feedback on quizzes ought to indicate why the choice(s) made were incorrect

to improve learnability.

Question 3 - As new features/exercises in PLOTs students requested for more quiz questions

and they also expressed the desire to be able to compare results with peers.

Besides the focus group open questions, students also filled an on-line survey, summarised in

Figure 1. The survey consisted of following 12 Likert-scale positive statements, which received

favourable responses:

Q1: The learning objects referred very well to the topic of the course.

Q2: The learning objects explained the topic in different contexts or situations.

Q3: The learning objects offered learning content beyond my usual e-learning expectations.

Q4: The content of learning objects provided appropriate examples to illustrate key points

or concepts.

Q5: I never wasted time with irrelevant details.

Q6: The learning content was easy to understand and concise.

Q7: Quizzes were at appropriate difficulty level.

Q8: The learning objects were easy to use.

Q9: The learning objects helped me identify my lack of knowledge.

Q10: The feedback I received for correct or wrong answers in quizzes helped me learn.

Q11: The amount of interactivity in the learning objects was appropriate.


Figure 1: Frequencies of answers in Likert-scale questions in pilot survey.

12

10

8 Unanswered

Strongly disagree

6 Disagree

Undecided

4 Agree

Strongly agree 2

0

Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11

Discussion and Conclusion In this paper we explored how persuasive learning designs can be effectively applied to the

design of PLOTs within a blended context capitalising the affordances of the VLE, with the

aim of changing students´ behaviour towards adopting a deeper learning approach. Filtering

key concepts into concise chunks, in clear and unambiguous language can provide more

targeted learning content with the persuasive principle of reduction. Students can clarify

assumptions and articulate knowledge interactively via the asynchronous discussions (the

principle of suggestion and social signals), and this in turn encourages students to take on a

deeper approach.

The findings evidenced that the PLOTs were well received by students and persuaded them to

engage with the learning and assessment activities independently. In summary, the content,

media, visual design allowed for persuasion to occur. The resultant PLOTs will be refined and

shared with the project’s partner institutions. Up until the time of writing, the study of

development of online learning incorporating persuasive design principles has yet to be

undertaken in Europe. Tashakkori and Teddlie (2010) state that studies involving human

behaviour change over time and are complex, therefore necessitating a sound understanding of

the social/behavioural phenomena that takes place. This merits further exploration and study

with a larger set of students to derive more conclusive explanations of the change in learning

behaviours instigated by the application of appropriate persuasive design principles specifically

through digital technologies.

Acknowledgements Part of this work was funded by the Education, Audiovisual and Culture Executive Agency

(EACEA) of the European Commission in the project EuroPLOT. We would like to thank the

coordinator and the partners for fruitful cooperation.


References

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W.H. Freeman and Company.

Biggs, J. (2003). Aligning teaching and assessing to course objectives. Teaching and Learning in Higher Education: New Trends and Innovations, 13-17.

Biggs, J., & Tang, C. (2007). Teaching for Quality Learning at University. Berkshire: Open University Press.

Clark, R. C., & Mayer, R. E. (2011). E-learning and the science of instruction: Proven guidelines for consumers

and designers of multimedia learning. San Francisco: Pfeiffer.

Fogg, B. J. (1999). Persuasive technologies. Communications of the ACM. 42(5), 26–29.

Fogg, B. J. (2003). Persuasive Technology, Using Computers to change what we Think and Do. San Francisco:

Morgan Kaufmann Publishers.

Fogg, B. J., Cuellar, G., & Danielson, D. (2003). Motivating, influencing, and persuading users. The Human-

Computer Interaction Handbook, 358-370.

Fogg, B. J., & Hreha, J. (2010). Behavior wizard: A method for matching target behaviors with solutions. In:

Persuasive Technology (pp. 117-131). Springer Berlin Heidelberg.

Fogg, B. J., & Tseng, H. (1999). The elements of computer credibility. Proceedings of the Conference on Human

Factors and Computing Systems, CHI99. Pittsburgh, PA USA.

Garrison, D. R., & Kanuka, H. (2004). Blended learning: Uncovering its transformative potential in higher

education. The internet and higher education, 7(2), 95-105.

Gram-Hansen, S. B. (2010). Persuasive Everyware-Possibilities and Limitations. WMSCI. IIIS, Orlando.

Gram-Hansen, S. B. (2012). D.3.3 PLOT Persuasive Learning Design Framework: Persuasive Learning Designs.

In: Persuasive Learning Objects and Technologies for Lifelong Learning in Europe.

Knight, J. (2010). Distinguishing the learning approaches adopted by undergraduates in their use of online

resources. Active Learning in Higher Education, 11(1), 67-76.

McLoughlin, C., & Lee, M. J. (2010). Personalised and self regulated learning in the Web 2.0 era: International

exemplars of innovative pedagogy using social software. Australasian Journal of Educational Technology,

26(1), 28-43.

Nicol, D. J., & Macfarlane‐Dick, D. (2006). Formative assessment and self‐regulated learning: A model and seven principles of good feedback practice. Studies in higher education, 31(2), 199-218.

Tashakkori, A. & Teddlie, C. (2010). Putting the human back in “human research methodology: The researcher in

mixed methods research. Journal of Mixed Methods Research, 10(4), 271-277.

Tseng, S., & Fogg, B. J. (1999). Credibility and computing technology. Communications of the ACM, 42(5), 39-

44.

van Dinther, M., Dochy, F., & Segers, M. (2011). Factors affecting students’ self-efficacy in higher education.

Educational Research Review, 6(2), 95-108.


Mediating Kaj Munk

Using Persuasive Learning

Christian Grund Sørensen, Sandra Burri Gram Hansen, Peter Øhrstrøm

Aalborg University

E-mails: [email protected]; [email protected]; [email protected]

Abstract. As part of the EuroPLOT project (PLOT = Persuasive Learning Ob-

jects and Technologies) the Kaj Munk Case is focused on disseminating

knowledge of the person, the works and the ideas of Danish vicar and play-

wright Kaj Munk. Due to the complexity of the content to be disseminated this

requires a multifaceted learning approach as well as support for existential re-

flection. Technologies used are The Kaj Munk Study Edition, geocaching and

mobile persuasive learning objects, and The Conceptual Pond. Combining these

approaches in the Munk Case provides a framework for technology enhanced

learning in the field of cultural mediation.

Keywords: Persuasive Design, persuasive technology, Kaj Munk, case studies

1 Background

In the course of developing the overall EuroPLOT-project the Kaj Munk Case was

suggested as an interesting and challenging case for the development and implementa-

tion of principles and techniques for technology enhanced learning using ideas from

persuasive technology. The Kaj Munk Case is developed by Aalborg University in

cooperation with The Kaj Munk Research Center, Aalborg, The Kaj Munk Vicarage

Museum, Vedersø, and several other institutions related to primary school, secondary

school, church education, folk high school and university.

Danish theologian, playwright, journalist and novelist Kaj Munk (1898-1944) dis-

played a comprehensive and widespread cultural impact in Denmark.

First, he is well known for a number of theatrical plays dealing with important reli-

gious, political and ethical questions. Being originally attracted to fascist ideas and

Moussolini’s ability to get the trains to run on time, Munk gradually rejected fascism

in general and the German Nazi regime in particular. He wrote several plays treating

these issues, acknowledging the importance of the strong character – yet reacting on

the progroms of the thirties. Other plays treated modernity, Christianity, faith, and

love.

Second, as a novelist Munk elaborated some of his existential thoughts in fiction,

© Springer-Verlag Berlin Heidelberg 2013


but in journalistic work for Danish newspaper Jyllands Posten he proved ardent and

sharp in rejecting the laissez faire approach to the rise of the Nazi regime and the

subsequent occupation of Denmark in 1944.

Third, as a clergyman and a preacher Munk sought to encapsulate his rather con-

sistent Christian reflections with ideas of war and resistance.

This being said, the fate of Kaj Munk is also decisive for the impact history of his

works and ideas. He was executed by German authorities in 1944, creating a myth as

one of the few Danish resistance martyrs. His (partially) nonviolent approach to re-

sistance and his respect for the common soldier amongst the enemy made him an

inspirational person for the post-apartheid peace and reconciliation process in South

Africa. So Kaj Munk was evidently a multifaceted personality with a vast range of

ideas and well-argued opinions adding to a notable life course.

It has been argued that Persuasive Technology as presented in (B.J. Fogg 2003)

should be seen together with elements of classical rhetoric (Hasle 2006). This is also

very much the idea on which the Kaj Munk case in EuroPLOT is based. One ambition

in the project has been to develop software which could on support the work of vari-

ous rhetors who want to guide the user. Another important ambition has been to make

it possible for the users to explore Kaj Munk’s conceptual universe by themselves.

The Emdros-based Kaj Munk Study Edition has as a software product been design to

function in both ways, being a tool for the rhetor who wants to guide the user and at

the same time being a media for the users who what to explore Munk’s ideas.

2 The complexity of the Kaj Munk Case

Each of the different cases in the EuroPLOT-project present different challenges to

the understanding of technology enhanced learning and the facilitation of multimedia

learning. Each case study displays a certain degree of complexity in the learning envi-

ronment, the subject matter, and the intended learning outcomes. Recognizing this, it

may be argued that in several ways the Munk Case does present an even higher level

of complexity than is expected in many generic learning situations.

It has thus been important to address the issues around this enhanced complexity in

order to achieve a thorough understanding of:

The communicative agent.

The intended learning outcomes or the intended persuasive impact.

The learning environment and recipients.

The material and subject to be mediated.

Obviously, these questions do apply in the process of designing most technology

enhanced learning systems. Nevertheless, inherent qualities in the material and the

agenda of the Munk Case prove considerations in this area to be of a most decisive

nature.


2.1 The communicative agent

In the Munk Case the potential transmitters of content and learning material are

legion. Such are also the intentions and the intended learning outcomes or intended

persuasive impact. In classic rhetorical theory the rhetor defines the agenda and the

arguments and, doing this, he enters into an act of rhetoric agency.

In the Munk Case there are many stakeholders and interests in the mediation and

learning situation. It has therefore been necessary to adopt an open approach to medi-

ating Munk offering tools and media for the users own, individual learning rather than

supporting a very fixed curriculum or a predestined route of digital immersion.

Communicative agents in the mediation of Munk could be teachers at different lev-

els of the school system. It could be University teachers, researchers, public opinion

makers. But it could also be the Kaj Munk Vicarage Museum or another cultural in-

tuition, or the Evangelical Lutheran Church. This diversity of potential rhetors obvi-

ously differs from much technology enhanced learning in a more restricted or struc-

tured learning environment.

In some cases the concept of a single communicative agent must be abandoned as

the learning environment is shaped by a multitude of users. This is for instance the

case where annotations or user generated content establishes any user as a potential

content provider.

2.2 The intended learning outcomes or the intended persuasive impact

Following the argument that multiple potential communicative agents are into play

in the Munk Case it is obvious that that a considerable degree of diversity will also be

present in the learning environment. Since rhetors have different agendas, motivation,

and intentions their focus in mediating Munk may be quite diverse.

Nevertheless, this diversity is not only a matter of intention; it is also a matter of

style and the role of persuasion. Thus, the communicative agent could aim at one (or

both) of these categories of recipient impact:

Intended learning outcomes – which should focus on knowledge dissemination and

cognition.

Intended persuasive impact – which should focus on bringing inspiration and even

suggesting attitude and/or behavior change.

In a habermasian understanding of learning obviously the intended persuasive im-

pact approach is ethically dubious. Nevertheless it is in cultural, political, or religious

mediation often the case, that mediation contains some kind of value transfer. It is the

strength of the Persuasive Technology framework (B.J.Fogg 2003) that it recognizes

these intentions and offer a theoretical framework for the discourse.


2.3 The learning environment and recipients

Just as the rhetors are diverse so are the receivers. Given that the learning environ-

ments may be at any part of the school system, at University, in Church or in cultural

institutions of different natures it is obvious that learning environments differ vastly.

2.4 The Kaj Munk Study Edition

Researchers at the Kaj Munk Research Center at Aalborg University have devel-

oped an Emdros based system which integrates archive-based learning resources (pic-

tures and texts), called the Kaj Munk Study Edition (Sandborg-Petersen, U. 2013).

This system exists presently in a Danish version. In addition, a German version and

an English version of this Study Edition are being developed based on the available

translations of Kaj Munk’s texts. This system makes it possible to search for patterns

and expressions in the text corpus. In this way the system offers an opportunity to

explore Kaj Munk’s universe as it is represented in terms of the texts.

The ambition is that The Kaj Munk Study Edition should include all parts of Kaj

Munk’s journalism, dramas, poetry, prose, sermons, and letters. In this way the work

with the development of the system may be taken as an illustration of how the com-

plete works of a writer of cultural significance may be represented using an Emdros

based tool and how such a tool may be helpful in order to solve tasks in a teaching

context, particularly the demands for problem-based learning. The ambition is also

that the system should be persuasive in the sense that the use of it should stimulate the

wish to learn more about the life and works of Kaj Munk.

2.5 Geocaching and mobile persuasive learning objects

In addition to the more traditional use of technology employed in the Kaj Munk

Study Edition, EuroPLOT has also made use of geocaching which may be seen as

special kind of so-called mobile persuasion. Within the project a system of four geo-

caches related to Kaj Munk has been established in co-operation with the Kaj Munk

Museum in Vedersø, which is located in Kaj Munks former vicarage. Two of the four

caches have been presented in both Danish and German. The four caches were in-

stalled in June 2011 and the use of the system has been studied over a period of two

years. There is strong evidence to support the claim that geocaching can in fact serve

as an effective persuasive technology.

Promising feedback from the Geocaching system in Vedersø, has lead to the de-

velopment of multimodal persuasive learning design, which was implemented and

tested in the small village Vester Hassing in Northern Jutland in June 2013. Building

upon the notion that the presence at a specific location, influences the users percep-

tion of the learning material, students at Vester Hassing school were introduced to

WWII and Kaj Munk, through a combination of traditional and mobile persuasive

learning objects. The learning design included a tour of the village, during which the


students received learning material and challenges at specific locations. For instance,

the students learned about the time when the railway in Vester Hassing was blown up

by the resistance during WWII, and challenged to capture a picture of the station in

the same angle as a picture taken when the station was still active.

Technologically, the learning design was based on the persuasive learning technol-

ogy PLOTMaker, which enables teachers to develop interactive and engaging learn-

ing objects that are executable either on computers or mobile devices. The mobile

GLOs hold particular potential compared to other similar systems, as learning materi-

al is not pushed from servers, but placed on the mobile device and activated at the

appropriate moment. Thereby, learning institutions are enabled to apply mobile learn-

ing material, without having to consider the cost of data streaming.

Observations from Vester Hassing showed that the learning design in general, and

the mobile learning design in particular, motivated the students to engage in the learn-

ing material in a highly collaborative manner. As the student walked between the

different locations in Vester Hassing, they would discuss the material they has just

been presented with, and “translate” it so that it was understood by all members of the

group. Furthermore, the evaluation of the learning experience, showed that the con-

structive approach to learning which forms the pedagogical foundation of the persua-

sive learning design framework (Gram-Hansen, 2012), resulted in all students meet-

ing the intended learning outcomes of the event. Finally, a majority of the students

responded that the learning experience had motivated their interest to learn more

about Danish and local history.

2.6 The material and subject to be mediated

As mentioned above the works of Kaj Munk contain much written material in the

shape of novels, sermons, plays and newspaper articles. Also a number of other

sources for the mediation are relevant: Contemporary photographs, records, discus-

sions of the subjects treated by Munk, extending the scope of Munk by interpreting

his beliefs and implementing them in a post-modern paradigm.

Accepting this means that the challenge of the Munk Case is to mediate a narrative

that is still in the making. This differs from the mere mediation of knowledge and

skill most akin to the general educational systems.

2.7 Complexity indesigned

In the Munk Case it has been a main concern to facilitate an approach to learning

that accepts the luhmanian autopoiesis and self-referentiality of the different subsys-

tems that are intended to make use of the learning tools and strategies of Europlot. In

order for this to be achieved an open strategy has been adopted in the Munk Case

offering ample freedom for the user to utilize the tools in an undetermined logic thus,

conceptually, developing an individual narrative in learning.

Having this focus on the different aspects of complexity in mind a number of tech-

nology enhanced learning applications have been implemented and tested.


3 Further research

In the course of development and testing in the Munk Case it has become obvious

that cultural mediation and the dissemination of ideas and relevant discussion is aided

by the establishment of multifaceted learning system. This system may consist of

several different applications and tools that may be implemented at the relevant time,

place and manner as suggested in the classic theory of kairos.

Two

3.1 The Conceptual Pond

In the course of the Kaj Munk Case an application was designed for representing

“snapshots” of the reception and understanding of a text, the so-called Conceptual

Pond (Sørensen & Sørensen 2013). The Conceptual Pond has been implemented for

the evaluation of the reception of Kaj Munks writings. So far mainly plays have been

used:

A. Kaj Munk’s play “The Word” in the version of renowned filmmaker

Carl.Th.Dreyer presented to confirmation class students at 13/14 years of age

and

B. Kaj Munk’s play “Love” presented to university students of applied philoso-

phy at Aalborg University.

In case A) the findings of The Conceptual Pond assessment system secured unique

knowledge revealing the reception and intuitive response to an unfamiliar style of

content related to complex existential and metaphysical questions. This is ground-

breaking as such research has not previously been performed for recipients of this

age. It may be argued that the intuitive and interactive assessment design of The Con-

ceptual Pond was of considerable help in A) acquiring the student’s responses and B)

in the process of providing statistical material for further research.

Though The Conceptual Pond offers valuable insights the concept could benefit

from further testing and development – especially in the field of enhancing the intui-

tive interface design and supporting expressions of opinion.

3.2 Immersive Layers Design

In the development of the different learning tools it has become clear that the

learning processes in the Munk Case could benefit from an enhanced categorization

and navigation system for the handling of the complexity of the subject.

The Immersive Layers Design consists of three separate layers in the interface that

function as different portals to a variety of digital learning materials. The three layers

are a geographical layer, a temporal layer, and a conceptual layer. The term immer-

sion is inspired by Murray (Murray 1997), pointing to deeper level of reflection.


Working with the conceptual approach is inspired by The Conceptual Pond (Søren-

sen & Sørensen 2013). The design is expected to be generalizable and applicable in a

number of technology enhanced learning environments of some similarity with the

Munk Case.

The threefold structure of the Immersive Layers Design is inspired by the rhetori-

cal concept of kairos. Acknowledging three aspects of kairos:

1. The opportune moment: Content is mediated at the contextually appropriate time.

2. The opportune location: Content is presented in suitable (virtual) geographical

surroundings.

3. The opportune manner: Content is mediated through the appropriate technology

(or the omission of technology).

Table 1. Immersive layers overview

Immersive layers

design

Layer 1 Layer 2 Layer 3

Description Geographical Temporal chronological

Conceptual

General properties

Visual, maps, pictures

Timeline, time shudder

Colour/picture based

Persuasive properties

Explorability, virtual locationbased,

reducing level of

abstraction,

Explorability, simulation, interac-

tion,

Intuitive access, overview, reduc-

tion, handling of

complexity

Learning value Reducing level of abstraction, acquir-

ing understanding

of geographical

context and real-

life events

Recognizing causal events, exploring

timelines, reference

to world outside

the Munk-universe

Acquiring overview, supporting

reflection, facilitat-

ing deep access to

knowledge systems,

linking content to

other layers

The Immersive Layers Design needs prototyping and testing for assessing the po-

tential in relation to especially the complex cultural mediation of the Munk Case.

4 Conclusion

In the Kaj Munk Case several applications have been implemented and tested. At

the same time the process has revealed that the complexity of the case seems to re-

quire an equally complex approach to facilitating technology enhanced learning.

Accordingly the different applications related to the Munk case should be under-

stood as parts of a comprehensive digital learning environment facilitating persuasive,

interactive learning in a manner relevant to the complexity og the case.


References

Burri Gram Hansen, S., Burri Gram Hansen, L., Grund Sørensen, C., Øhrstrøm, P.:

PLOTLearner v3: The Kaj Munk Case, Deliverable 5.3. EuroPLOT 2013.

Burri, Gram Hansen, S., Kristensen, K.Dyrby, Burri, Gram Hansen, L.: Motivating the In-

terest in Danish Literature with Mobile Persuasive Learning, IWEPLET 2013

Fogg, BJ: Persuasive Technology - Using Computers to Change What We Think and Do.

2003, San Francisco: Morgan Kaufmann Publishers (2003)

EuroPLOT: Persuasive Learning Objects and Technologies. http://www.eplot.eu/ (2010)

Hasle.P., Christensen, A.K.K (2007): Classical Rhetoric and a Limit to Persuasion. Persua-

sive 2007, LNCS 4744. Springer

Hasle, P. (2006) The Persuasive Expansion - Rhetoric, Information Architecture, and Con-

ceptual Structure, in ICCS 2006, H. Schärfe, P. Hitzler, and P. Øhrstrøm, Editors. 2006,

Springer

Murray, Janet (1997) Hamlet on the Holodeck: The Future of Narrative in Cyberspace,

New York

Sandborg-Petersen, U. Architecture of applications built on Emdros: Case Studies in Sys-

tems for Persuasive Learning, IWEPLET 2013

Sørensen, C.Grund & Grund Sørensen, M. (2013) The Conceptual Pond – a Persuasive

Tool for Quantifiable Qualitative Assessment in “Emerging Research and Trends in Inter-

activity and the Human-Compute Interface”, IGI Global

Sørensen, C.Grund: Intuitive Surveying & Quantification of Qualitative Input through the

Conceptual Pond, IWEPLET 2013

Sørensen, C.Grund: Immersive Layers Design Exploring Culture through a Persuasive

Multimodal Interface, IWEPLET 2013

1. Gram-Hansen, Sandra Burri. (2012). Deliverables Report D.3.3, Persuasive Learning

Designs: Aalborg university.


PLOTLearner for a Corpus of the Hebrew Bible: The Case for Repurposing in Language Learning

Nicolai Winther-Nielsen Fjellhaug International University College Denmark,

Frederiksborggade 1.B.1, DK-1360 Copenhagen K [email protected]

The Lifelong learning project EuroPLOT (www.eplot.eu) has developed a learning

technology called PLOTLearner that applies principles formulated in the theory of

Persuasive Technology. This new persuasive technology uses a text database to drive a new

kind of language learning that focuses on enhancing motivation and enablement through

persuasive principles. It also focuses on repurposing learning objects in a persuasive

context. The database is the corpus of the Hebrew Bible from the Eep Talstra Centre for

Bible and Computer at the Vrije Universiteit in Amsterdam. The interface to this learning

technology and the content and structure of learning objects were tested in Gothenburg and

Copenhagen and then repurposed as a sustainable technology in Madagascar. The Hebrew

test case demonstrates how the technology scales persuasively from small test groups to

500 students in a situation where a national institution in Africa sought to apply the

technology in a way that complies with the Bologna process.

Keywords: Learning technology, Design for sustainable language learning. Corpus-driven

language learning, Hebrew Bible database. Madagascar.

1 Introduction EuroPLOT develops Persuasive Learning Objects and Technologies, or PLOTs for short, for

reuse and repurposing in several learning contexts of language, culture, environment, and

business IT. This case study for language learning uses the corpus of the Hebrew Bible in a

persuasive technology called PLOTLearner to demonstrate transnational repurposing for

tertiary students at the university in Gothenburg and at a college in Copenhagen and then scaling

up to vocational and tertiary learners in Madagascar.

Implementing learning technology in Madagascar is a challenge because it ranks as one of the

poorest countries in the world, and the use of computers is not common. Nevertheless, focus

interviews fortunately indicated how motivational this learning technology is, as in the

following excerpt of an interview with a teacher training to repurpose the technology for his

own teaching:

When we have PLOT Learner, it is like as we have seen a light, because it is very

clear how to teach Hebrew with PLOT Learner. Also, the way to persuade the

students to learn Hebrew through using the computer is very interesting. In the area

where we are, 70% of the students have never used a computer. But PLOTLearner

has helped them … learn how to use a computer, and they are motivated to do so in

order not stay computer-illiterate. The lessons in the PLOT Learner are very simple

and organized. They didn’t like Hebrew classes before, and many students gave up.

But now they compete in doing exercises with the software. (Pastor

TSANGAMILA Fils, Hebrew teacher from STPL Vangaindrano,

http://bh.3bmoodle.dk/mod/resource/view.php?id=221, 1:40-4:44.).

This paper will discuss the implementation of persuasive learning technology and especially

focus on the goal of the project to effectively use “principles for designing technologies that


change behaviour or attitude, for example, by using fun, simulation, competition, cooperation,

or peer influence” (www.eplot.eu). After an introduction to the history and vision of the case

for Hebrew language learning, the presentation explains results from the initial pilot testing of

the tool for Hebrew in Gothenburg and the creation of content for classroom teaching in

Copenhagen. The main focus is to explain how the scaling up for Madagascar has succeeded in

repurposing the technology and training facilitators who will teach some 500 students annually.

This is followed by a section explaining how data was collected.

2 PLOTLearner - technology for a corpus of the Hebrew Bible The EuroPLOT application in 2010 formulated the novel idea that a database application

PLOTLearner could turn a linguistic corpus like the Hebrew Bible into a pedagogical tutor of

the learner. Using the core concepts of reusable and repurposable learning objects, the project

envisioned effective learning through interactive and pedagogical engagement with digital

learning objects generated by the Hebrew texts. Learners would gradually be persuaded to

improve their learning attitudes and activities, while being gently guided by the technology to

learn from the texts and practice language skills.

The story that informs this novel idea began with experimental use of technology in the Hebrew

classroom back in 2003 (Winther-Nielsen 2011). In the classroom there were good results with

a simple program, Paradigms Master Pro (http://paradigmsmasterpro.com/), which generated quizzes

on nouns, verbs, suffixes, and prepositions. Not only did drills help students learn much faster

and often with fun in competition, but some would even develop an ability to recognize

structure in a fashion similar to a parser. Students in this classroom would use text, grammar,

and dictionary resources produced by the commercial Bible software company Logos

(https://www.logos.com/). Among those resources was a rich linguistic database of the Hebrew

Bible developed since 1977 by the Werkgroep Informatica at the Vrije Universiteit in

Amsterdam (WIVU, now renamed as the Eep Talstra Centre for Bible and Computer).

This data is stored in an Emdros datatabase management system (www.emdros.org). The

principles embedded in this database were developed by Ulrik Sandborg-Petersen (2011) of

Aalborg University into a database for storage and retrieval of data from text databases. This

database system was in turn used for linguistic projects on Role and Reference Grammar, and

it led Winther-Nielsen (2011) to formulate the vision that the linguistic corpus of Biblical

Hebrew could be used for persuasive language learning. From 2008 we experimented with

combining the effects of drills observed in the classroom with database technology. Then in

2010 Claus Tøndering wrote and released the Windows application 3ET, short for Ezer Emdros-

based Exercise Tool, which generated quizzes from the Hebrew corpus. The next step was to

develop the idea of a PLOTLearner, which would add persuasive learning functions to 3ET in

the way described in Winther-Nielsen (submitted; ms b) and Skovenborg (2011).

EuroPLOT disseminates PLOTLearner for open global educational use in Hebrew language

learning. The leading publisher of academic editions of the Hebrew Bible, the German Bible

Society in Stuttgart, has granted EuroPLOT a license to offer 20 % of the Hebrew database for

free. Therefore, there are no obstacles to widespread distribution of PLOTLearner due to cost

or quality of content. The challenge of the project is rather to substantiate the claim that this

new approach to Hebrew language learning is superior to other alternatives, which is the aim

for this discussion of PLOTLearner in terms of pilot testing, scaling and solid statistical data.

3 Piloting PLOTLearner in Gothenburg and Copenhagen


PLOTLearner has from the outset been developed through participatory design focusing on

feedback from user experience.

The first main prototype of the Windows program, PLOTLearner 1, was released in October

2012 and had enhanced support for transliteration and user-friendly presentation of exercises.

This version was tested by 6 students of EuroPLOT associate partner Nava Bergman from

Gothenburg University in February 2012 through quantitative surveys. She invited some of her

former Hebrew students to participate. 9 students began the test. 6 students completed it

(Winther-Nielsen 2012). Even if this group was small and not randomly selected, it served well

as a heuristic pilot testing group. Participants were asked to answer a pre-test questionnaire,

then test the tool for 10 hours before doing a post-test questionnaire. The evaluation measured

the attitudes of the respondents, supplemented by optional comments.

This evaluation focused on the exercise support of the program, which got the highest scores

along with training of morphology. While respondents noted insufficient learning content, they

clearly sensed the persuasive effect of the program. Among the helpful responses received were:

it is brilliant to do text-based exercises; it is very helpful to have a virtual keyboard for typing

Hebrew; it is very useful to have access to mouse-over popup information from the database.

Crucially, 3 students found PLOTLearner clearly motivational, and only one disagreed.

Respondents pointed out that the program was fun to use, increased desire to practice, and was

effective when one got used to the tool and to learning at their own pace and at any time they

liked.

However, this test group knew Hebrew and did not need a tool for beginners and therefore did

not want to participate in further testing. Furthermore, the university discontinued teaching of

Biblical Hebrew, and therefore Bergman could not scale up to the planned 40 students for the

final test. Currently teaching assistant Christian Højgaard is using EuroPLOT’s PLOTMaker to

repurpose learning objects for the textbook of Bergman (2005), and hopefully this will attract

students eventually.

Two students in Copenhagen participated in Think-Aloud Observations recorded on video in

March. The protocols and recordings confirmed the positive evaluation from Sweden. Direct

observation in a controlled environment enabled tracking of persuasive learning and the

observer could gently guide the informant towards issues of interest (Rogers et al 2011: 256).

The second prototype, PLOTLearner 2, was developed into a simulation of a corpus-driven

study of the Hebrew text and supported learner-friendly interaction. This version was released

July 26, 2012. During the Autumn of 2012 EuroPLOT produced and tested 12 sessions for

introductory teaching of Biblical Hebrew (http://bh.3bmoodle.dk/course/view.php?id=2 (login

as quest)). The reporting on the tests results from this prototype is still in progress (Winther-

Nielsen ms a).

The aim of this phase of the EuroPLOT project was to produce teaching material for

experimentation with Hebrew language teaching as blended learning and to test persuasive

learning objects. The 12 sessions were tested during the first two-and-a-half weeks of an 80-

lessons-long course on Biblical Hebrew given at Fjellhaug International University College

Denmark, a campus accredited by a Norwegian college. It was therefore possible to test the use

of PLOTLearner by 5 students over a period of four months until their final exam.

Exercises and learning material was delivered in Moodle. The students primarily used short

videos introducing grammatical topics as well as one-page pdf-sheets with condensed

overviews on these topics. All material could be accessed through a text in PLOTLearner

describing the exercise. Class sessions were recorded in Camtasia screen capture, and Skovgård


(2012) analyzed the relation between teaching and outcomes achieved in PLOTLearner, while

Gottschalk (2012) explored the users’ interest in support for collaboration.

Unfortunately, EuroPLOT had no clear immediate positive influence on the final exam

performance for these students. They did not do as well as was expected in the language exam,

although they performed well in translation. Their eventual performance was good: half a year

later they all got A and B levels in a subsequent course on interpretation of texts from the

Hebrew Bible. The lesson from this experience is that learners must be tested in accordance

with the skills PLOTLearner builds in acquisition of morphology, vocabulary, and translation.

In an ideal world the often superfluous final exam should be replaced by regular or at least

weekly self-directed tests for documentation of individual progress. This can develop skills

during a learning process in line with the new learning promoted by the European Union.

Our most important conclusion from this pilot evaluation of the two prototypes and the

participatory development of learning content is that PLOTLearner integrates well into a

classroom with blended learning, supporting skill acquisition in morphology, vocabulary, and

translation. However, facilitators of PLOTLearner must repurpose learning material to match

grading requirements, if they cannot change evaluation and testing mechanisms to match and

support learning as a process.

4 Malagassy education and teacher training Due to the discontinuation of Bergman’s teaching of Biblical Hebrew at Gothenburg

University, EuroPLOT has only had the opportunity to scale PLOTLearner for its final pilot

testing in Madagascar in the first quarter of 2013.

The technology is implemented to change the teaching of the Hebrew Bible in an entire

educational system for vocational training of pastors in the Malagasy Lutheran Church. This

church has an estimated 3.2 million members and urgent needs for an increase in educational

capacity. Many pastors in rural areas serve more than 10 churches, covering a large

geographical area with poor infrastructure, and the churches are growing fast. An average of

400 students are enrolled in a 4 year pastors’ training program at six rural pastoral training

schools, and close to 100 students in a 3 year program at the Lutheran Graduate School of

Theology, or in short SALT, which is involved in EuroPLOT as an associate partner.

The Malagasy representative of the EuroPLOT associate partner, Dean of SALT, Dr Lotera

Fabien, wanted PLOTLearner to enable students training at the vocational STPL schools to

qualify for admission in the advanced bachelor education at SALT. After two years of study at

the STPL, they apply for finishing the last two years of their four-year Bachelor’s program and

a three-year Master’s program at SALT. Many of these candidates will get teaching jobs while

others will go into administration or become lead pastors. This is part of the plan to upgrade

education under the Bologne process. EuroPLOT will provide the new kind of learning in

demand locally.

On behalf of EuroPLOT, teaching assistant Andrianotahina Naivoson Hery directed the testing

of PLOTLearner 1 with a group of 15 students beginning January 2012. From September to

November 2012 he installed PLOTLearner 2 at each STPL in order to help Hebrew teachers to

get a first personal experience with the technology. The fieldwork for pilot testing of

PLOTlearner by this author was scheduled from January to March 2013.

The first job was to enhance the professional training for 5 out of 6 teachers of Hebrew at the

rural STPL schools. They both needed to learn more Hebrew from the 12 EuroPLOT sessions

and more advanced content. They also needed to learn how to run tests in PLOTLearner and


provide documentation of skills of their students. The challenge was to transform teaching

experience documented in the classroom in Copenhagen into a contextualized training for a

new facilitator role scaled to the needs in Madagascar. This teacher training was carried out in

an intensive 8-day-long introductory course on PLOTLearner in January 2013. Six weeks later,

after the 5 Hebrew teachers had experimented with self-directed learning, they returned for a

final course week. In this phase, Malagasy teaching assistants gathered invaluable feedback in

questionnaires and video recorded interviews.

Repurposing of educational material in Madagascar has to adjust to a situation without fast and

affordable internet access. Delivery in Moodle is impossible, but one-page pdf-sheets with

condensed grammar work well. All resources were now built around a Word97-2003 document

integrating hyperlinks to PLOTLearner, videos, and pdf documents. Teaching assistant Tiana

Andriamapandry (2013) used the 12 EuroPLOT sessions and existing material and repurposed

it all into Malagasy for the first pdf-edition of Boky Voalohany. This teaching grammar

integrates with PLOTLearner and will be used and tested at the STPLs and further developed

in response to Malagasy learner feedback.

The fieldwork included an evaluation of the learning environment in which the new

PLOTLearner was going to be used. It came as an unpleasant surprise that some 400 students

at 6 STPLs only had access to 2 computers! We estimated that basic needs could be covered if

they had 1 computer per 10 students. Thanks to donations to the amount of 3500 Euros, the

project can now distribute 40 second-hand, well-functioning Windows computers and 6 laptops.

This will provide access to learning content via USB-keys, but it will also promote collaborative

learning in groups and pairs.

For this new leaning environment we explored a new approach that would foster self-directed

learning, overcome one-way communication, and demotivate copying from the blackboard.

These teachers are now contextualizing learning as a bonded fellowship, inspired by the local

culture’s core concept of collaboration across age groups, known as fihavanana (Dahl

1998:139-143). This kind of networking and collaboration does not only apply to the extended

family, but can be used for any relationship based on respect, care, mutual favours, and trust,

which can form the core for a new kind of Malagassy learning 2.0.

In a report from STPL Atsimoniavoko on June 19 2013, Hebrew teacher A. Pascal

Ramahafadrahona describes how he has formed small groups for both his Hebrew classes, 19

groups for the junior class of 80 students and 5 groups for the senior class of 40 students. Even

if some students are shy or resist collaboration, the majority of the students help each other,

competing as a group and making good progress. He used the tool for the final Hebrew exam

at his school, working for an entire week to complete the exams on a few laptops borrowed

from students and friends in the nearest city. This teacher repurposed exercises from

PLOTLearner for the final test, and forwarded examples show impressive results from these

students.

In this way local teachers are now repurposing the technology into a truly sustainable learner-

centered PLOTLearner Fihavanana learning environment. It will work without unaffordable

textbooks from the West and it will support effective learning in the learners’ mother-tongue.

EuroPLOT’s persuasive technology is thus already being implemented for teaching of 500

students annually at the end of the project in October 2013. The feedback from evaluations

testifies to students having fun, completing their learning, and engaging in collaborative

learning.

The intensive teacher training course gave invaluable data on the persuasive force of

PLOTLearner for facilitators in a global setting. Whereas before students barely learned to read

the Hebrew text, students can now practice language learning skills in morphology, vocabulary,


and translation, and teachers and their assistants can easily repurpose the technology to local

demands. It enhances new persuasive learning 2.0 for sustainable effect.

5 Statistical data on learner progress The claims for persuasive technology made so far do not stand alone, because the deployment

of PLOTLearner in the Hebrew case study has focused on gathering solid statistical data on the

persuasive effect of this new learning technology. The evaluation of the technology and the

learning content in pilot groups and in a full scale implementation in Madagascar has provided

learner data as evidence on how technology can motivate learners to track their own progress

and improve their own learning journey. This illustrates also how teachers are able to survey

the progress of students and facilitate students in their learning project and improve the learning

outcomes for students.

Figure 1. The graphical and statistical displays of test results in PLOTLearner

PLOTLearner provides feedback on persuasive effect in learning outcome by displaying

statistics on learner performance. The evaluation has so far been based on the “Complete local

report”, which is shown in the right hand side of Figure 1, and the “Exercise graph” in the left

hand side of the figure (See Tøndering 2012: http://www.ezer.dk/3ETusersguide/PL-

2.0.1/en/statistics.php). These statistics files were used for manually gathering data on student

performance related to duration of exercise, seconds elapsed per right answer, and how many

right and wrong answers.

On February 8, 2013, a two-week course on the 12 sessions on introductory Hebrew was

concluded with a 15 minute long test for 79 students at SALT. This test has 104 questions in

four exercises covering the first five verses of Genesis chapter 1. In figure 2, the light blue

graph from the left lower corner starts at the 3 seconds per right answer (S/R Feb) for the best

student in this test. This graph slowly rises up to 30 seconds per right answer for the first 53

students, and then jumps steeply upwards, and 21 students, or every fourth student, failed to


complete the test. The red line also begins in the lower left corner at the 2 wrong answers (Wrg

Feb) for the best student. The variation in this curve on wrong answers illustrates that the second

per right value is not a perfect indication of best student score because some learners are fast

but sloppy and make many mistakes, while the best learners will be both fast and accurate.

Figure 2: Results from Tests at SALT in February and July.

A series of experimentations with learner statistics therefore made us change the calculation of

language proficiency when we had the Malagasy teaching assistants gather data for the same

test a second time by the end of the school year in July. We still gather the second per right

value (S/R July) illustrated by the green graph which now is a more or less flat curve ranging

from 3.19 to 7.32 seconds for 73 students out of 78 participants. These students had clearly

learned to master PLOTLearner during the intervening 5 month of practice before the second

test. We now used a new calculation of the average between right and wrong answers which

has been introduced for the Learning Journey Online (Gottschalk and Winther-Nielsen ms).

The average of the right answers is multiplied by three, and therefore this value begins with a

high value of 38 points for the 13 best learners who had no mistakes at all in this test. 66

students had values over 33 point which is equal to 17 mistakes and the cline only falls for the

last 12 students in this test ending at 26 wrong answers for the poorest student.

This data illustrates the progress among a statistically valid population, but it does not cover an

advanced knowledge of Hebrew. However, our experiments with data gathering not only proves

that we now have a very stable system, but also that we will be able to provide persuasive

feedback and self-monitoring as we continue to develop this kind of corpus-driven learning. In

our typical courses students would do this 15 minutes test after 12 sessions. A learner would

then prepare for a one-hour test on exercises generated by the corpus of Genesis 1-3, testing

knowledge on nouns for 10 minutes, verbs for 15 minutes, vocabulary for 10 minutes, and

translation for 20 minutes. After these first 30 or so sessions on introductory Hebrew, a learner

can pick any text and continue to explore and practice skills.

7 Conclusion

The presentation of the case for persuasive learning of Hebrew has focused on how learners

and facilitators can use PLOTLearner for motivation and enablement. The first step was to

explain how the technology was developed through pilot testing and improved in agile

programming. In the next step we summarized repurposing as open and sustainable technology

in Madagascar. The final step introduced our methods for gathering of statistical data that

corroborate our claim that this technology is effective for language learning.


From the pilot testing in Gothenburg we realized how important it is that persuasive technology

is user-friendly in its integration of learning content. The main result from Copenhagen is that

corpus-generated practice and feedback works in blended learning, but it must match the

requirements in the final exam. From scaling up to Madagascar we are able to plot

PLOTLearner’s potential for repurposing learning to an education by training teaching

assistants who will repurpose resources to the local culture. Finally, the statistical data on

learner progress and performance are helpful for tests and for monitoring of student progress.

This Hebrew case study presents a strong case for persuasive language learning driven by a

text-database and it could work for any text and any language.

Acknowledgement The EuroPLOT project was funded by the Education, Audiovisual and Culture Executive


grant #511633. I would like to thank Claus Tøndering and other partners, teaching assistants

and students mentioned in this paper. Randall Tan helped with language editing.

References

Andriamapandry, T. (2013). Boky Voalohany. Malagassy Hebrew Grammar, PDF.

Bergman, N. (2005). Cambridge Biblical Hebrew Workbook, Cambridge: Cambridge University Press

Dahl, Ø. (1998). Merkverdige Madagaskar – øya mellom øst og vest, Oslo: Spartacus Forlag AS,

Gottschalk, J. (2012). The Persuasive Tutor: a BDI Teaching Agent with Role and Reference Grammar

Language Interface. ITB Journal, 23, 31-51.

Gottschalk, J. & Winther-Nielsen, N. (ms). Persuasive Skill Development: On Computational Surveillance

Strategies for Modeling Learning Statistics with PLOTLearner. Submitted to IWEPLET13.

Rogers, Y., Sharp, H., & Preece, J. (2011). Interaction Design: Beyond Human Computer Interaction. 3nd Edition.

Chicester, West Sussex, UK: Wiley

Sandborg-Petersen, U. (2011). On Biblical Hebrew and Computer Science: Inspiration, Models, Tools, and Cross-

Fertilization. In W. T. van Peursen & J. W. Dyk (Eds.), Tradition and Innovation in Biblical Interpretation.

Studies Presented to Professor Eep Talstra on the Occasion of his Sixty-Fifth Birthday (pp. 261-276). Leiden:

Brill

Skovenborg, D. L. (2011). A new PLOT for Biblical Hebrew learning. Master Thesis, Aalborg University.

Retrieved July 26, 2013, from http://waldeinburg.dk/upload/Skovenborg%20%20A%20new%20PLOT%

20for%20Biblical%20Hebrew%20learning.pdf.

Skovgård, M. G. (2012). Læring med PLOTLearner- Et bud på en ny måde at undervise i døde sprog. Project

report, Aalborg University.

Tøndering, C. (2012). PLOTLearner User’s Guide and Reference. Retrieved July 26, 2013, from http://www.ezer.dk/3ETusersguide/ PL-2.0.1/en/.

Winther-Nielsen, N. (2011). Persuasive Hebrew Exercises: The Wit of Technology-Enhanced Language Learning.

In W. T. van Peursen & J. W. Dyk (Eds.), Tradition and Innovation in Biblical Interpretation. Studies

Presented to Professor Eep Talstra on the Occasion of his Sixty-Fifth Birthday (pp. 277-298). Leiden: Brill

Winther-Nielsen, N. (2012). Evaluation of PLOTLearner 1. Deliverable for EuroPLOT (available at

http://bh.3bmoodle.dk/mod/resource/view.php?id=113 Login as guest) Winther-Nielsen, N. (submitted). The Corpus as Tutor: Data-driven Persuasive Language Learning. Submitted to

Digital Humanities Quarterly.

Winther-Nielsen, N. (Ed.) (ms a). Evaluation of PLOTLearner 2. Deliverable for EuroPLOT. Draft in progress.

Winther-Nielsen, N. (ms b). PLOTLearner as Persuasive Technology: Tool, Simulation and Virtual World for

Language Learning. Submitted to IWEPLET 2013.


Designing The Persuasive Learning Experience

Erich Herber

Danube University Krems, Krems, Austria

[email protected]

While the persuasive strength of a digital learning object will vary from learner to learner

or from learning setting to learning setting, there are persuasive design strategies that are

likely to apply to a larger number of educational settings. This paper discusses some of

these design strategies and attempts (from the viewpoint of persuasive learning) as well

as design considerations that shall help designers of digital learning objects to develop

high-quality persuasive learning designs for their audiences.

The objective of this paper is to present key findings of research that has been made in

EuroPLOT, a 3-years LLP project, as well as to propose overall design recommendations

for persuasive learning objects. We are aiming at the involvement of persuasive strength

as a new principle when designing digital learning objects.

Keywords: Learning Design, Persuasive Design, Persuasive Learning, Evaluation

Introduction

This is a paper resulting from the evaluation of the EuroPLOT project and explores the

interrelations of the two disciplines learning design and persuasive design. It is based on

research undertaken among a sample of learners in four pilot projects of different educational

areas. The aim of this particular study was to gather evidence in relation to the following key

question: How should persuasive learning experiences be designed?

We started to elaborate on this question by approaching the user of learning objects as a

multiple character: The user is a learner (aspects of learning design) but also a user who

interacts with an interactive learning system (aspects of usability design) and responds to

emotional encouragement and involvement (persuasive abilities of design) throughout the

process of learning. We are aiming at the involvement of persuasive strength as a new

principle when designing and/or evaluating digital learning objects. We want to extend the

current practice of learning designs by focusing not only on cognitive but also on persuasive

considerations that relate to learning. Our approach goes beyond the functional usability

paradigm of learning design and proposes persuasiveness as a new type of measurement.

Methodology The underlying study of our paper was both quantitative and qualitative in approach and a

mixture of focus group discussions and online surveys was used. The first stage of this was a

series of in-depth discussions with learners who had worked with persuasive learning objects

that designers of the project developed. In addition to these discussions, an analysis of data

from an online questionnaire was conducted to gather further information about provision

across the three different contextual sectors in the study. Also a review of prior studies in the

disciplines of usability design, learning design and persuasive design was conducted to

identify complementary design strategies. With respect to the later, we particularly looked at


Persuasive Design as a design strategy that enables designers of learning objects to respond

more flexibly to motivational and effective needs of learners (Fogg 2003).

Key Findings In practice, there is a positive attitude towards the overall appearance and instructional

approach of the developed learning objects across all disciplines. The instructional designs of

the piloted learning objects were felt (by the learners) to be both creative and comprehensible,

with a high level of concern for active learner engagement and motivation. Constructive

alignment has been found to be a key principle for which learners perceived learning objects

as beneficial for their learning. However they expressed a need for more self-reflective forms

of learning and more elaborated interactive content. Particularly challenging learning

experiences with explorative learning activities were considered to be highly useful for their

learning, as well as more established components such as quizzes, progress control and step-

by-step guidance.

Both learners and designers of the EuroPLOT pilot projects generated debates about the

complexity of learning resources and the appropriate level of detail that was necessary to

present new concepts or ideas. They all agreed that there is a fine line between a too narrow

and a too broad presentation of details. Learners from the Higher Education sector (e.g. the

Hebrew learning case) critically questioned some of the exercises of the piloted learning

resources in the light of different curriculum expectation levels (e.g. between beginners or

advanced learners, lower or higher-grade levels) whereas learners from the Vocational

Education sector rather felt persuaded when learning activities would prepare them

sufficiently for practical work situations and related to practical contexts (e.g. in the Exposure

scenario case). Instructional design strategies that set value more on a learner specific

customization and/or reduction of the digital learning content will potentially lead to

contribute to the point of persuasion (see ‘Tailoring’ and ‘Reduction’ by Fogg 2003).

The most consistent lack across all the three disciplines was identified with respect to an

absence of self-monitoring activities as well as little stimulation towards immersion (see ‘Self-

Monitoring’ and ‘Suggestion’ by Fogg 2003). Some learners also expressed concerns about

the specific practices and relevancy of how feedback and response needs to be provided in the

digital learning experience. In fact, not all learners were aware of the entire range of

possibilities that the feedback options of the learning resources offered them to enhance their

learning. Hence, important educational and/or motivational effects have been lost. In some

cases, if feedback was of less immediate concern to a learner’s own attitudes or concern, the

feedback was not raised as utterly motivational or it was even ignored as such.

A key aspect of the designers’ role in the whole learning experience is about negotiating

commitments to learning as well as about assuming the positions and learning preferences of

the targeted learners in advance. Typical questions that designers with an attempt to

persuasive learning strategies should think about are: “Who are the learners and what is their

current opinion on the taught issue?”, “What might be the motivation or purpose for their

opinion or attitude?” or “What arguments or pedagogical strategies are most likely to

persuade them?” Despite the fact that knowing the audience is one of the key qualities of the

expert learning designer, designers should be clear of the fact that they will not be able to

persuade every single audience member with their learning designs!

While the persuasive tactics of learning designs will vary from occasion to occasion, there are

tactics that are most likely to apply to a larger number of learning scenarios or settings. The

following list presents a set of persuasive design consideration that may help designers to


embed persuasive tactics in their learning designs. These tactics (which are only exemplary at

this point) will go beyond common instructional design guidelines, but do not necessarily

compete with them:

Attention. The first impression is important! The opening of learning objects therefore is a

significant and opportune moment to capture attention of learners and/or to shape their

opinion. Contradictory statements, questions or humor can increase attention quite easily.

Time. Learners do not have time to waste with irrelevant details, clicks or links. If

designers wait too long with explaining and/or demonstrating key issues and concepts, or if

they are too importunate with too much repetition and immersion, the learning object can

lose people’s attention as well as their credibility.

Detail. The designer of persuasive learning content should make sure that the main points

are concise, clear and presented to the level of detail that is relevant to achieve the targeted

learning outcomes.

Efficiency. Smaller and less time consuming learning activities will make the performance

of a learning object more practicable and motivating for learners than complex learning

paths. The higher the complexity of a learning activity, the more time learners will have to

invest to work efficiently through the learning experience.

Reinforcement. Concluding statements or reflective tasks at the end of a learning object

(e.g. quizzes), or at a crucial point during the learning experience, can help to repeat core

concepts or positions without too much of redundancy or force.

Simplicity. Designers should make sure that interactive tasks do not require too much time

or efforts from the learner. Otherwise they may lose their attention, or distraction may

occur. When learners lose attention they are not likely to be persuaded.

Guidance. Designers of learning objects should also give unambiguous and specific

instructions of what learners are asked to do.

Benefits. When new opinions or positions are explained in the learning object, the benefits

and motivation of the position should be clear to the learner. From the designer’s point of

view, the benefits that are given should respond to the very position of the targeted

learners, their contexts, beliefs or attitudes.

Flexibility. Didactic flexibility and diversification is the best tactic of designers to make

that action request. If the target learner hesitates to accept a targeted learning activity, the

learning object should offer an alternative activity (for example a less complex task, a

different didactic approach, or a new learning journey).

Feedback. Learners like to know if their actions have made a positive difference. They

should be informed of the consequences of their commitment and/or action. Even if the

intention of the feedback is not primarily intended to motivate learners, this is how the

feedback is usually perceived by the learner.

Commitment. Any kind of committed action (e.g. accepting a button to start a learning

activity or to confirm or disagree with an opinion) will lead to attention and active

engagement in the learning experience, and potentially result in higher persuasion. Instructional designs that consider either of these tactics are more likely to answer to a high

persuasive design attempt of a designer. However, designers must understand that persuasive

learning objects are not persuasive simply for the fact that they embed persuasive strategies or

apply persuasive design principles (Fogg 2003). A core aspect of persuasive learning design is


the building on the interests of learners. Persuasive learning design is expected to clearly

identify benefits with the greatest appeal to the users of the learning objects. Designers are

requested to respond to them most flexibly for the purpose of persuasion when developing the

learning content. Also rethinking the channels and methodologies of self-regulated learning as

well as increased learner involvement in the learning experience will respond to the specific

demands of persuasive learning design.

With no doubt, learning motivation and attitudes are likely to benefit if flexible, personalised

and learner-centred learning strategies are embedded in learning designs. Some learners

stressed the fact that they benefited from the digital learning objects (as a complementary

learning resource for their face-to-face teaching units) because they were able to deepen their

knowledge and experience in a more explorative from, particularly when they studied new

and/or more complex concepts. The learning activities helped them to see new aspects of a

topic (or to reflect more thoroughly on them), or even lead to discover misconceptions. It was

obvious that particularly the interactive aspects of self-regulated learning had an impact on

learners’ motivation to learn as well as that different forms of interactive learning enhanced

their attraction and interest in the whole learning experience.

As regards the content and the intended learning experience of a persuasive design attempt,

the following characteristics were identified as a result of the different design considerations

that were raised in discussions with learners, designers as well as literature reviews (see Table

1).

Table 1. Evidence of Persuasive Learning Designs

Designing the Persuasive

Learning Content

Designing the Persuasive

Learning Experience

The content offers benefits or rewards

to the intended audience.

The content offers prompt, constructive

and personalised feedback.

The content suggests that a learning

benefit is particularly valuable or

scarce.

The content is mediated at the

contextually appropriate time.

The content appeals to the norms and/or

learning preferences of the learners.

Key concepts of the content are

repeated without too much force to support reasoning or to gain attention.

The content refers to other, more

complex material (on request).

The content is mediated through

appropriate technology, for example

interactive and explorative tasks.

Content is delivered in an interactive

and self-controlled learning pace.

The learning experience is consistent with

previous beliefs, actions or expectations of the

audience.

The learning experience has a time limit and/or

deadline associated with the experience.

Learners actively engage in the learning

experience and have autonomy on how to

complete learning activities.

The design elements, language, formatting,

appearance, and functionality are used in a

consistent form.

The design is simplified and sticks to basic

principles of aesthetics.

Learning activities enhance the self-reflective

skills of learners.

Feedback and control is part of the learning

experience, not just summative.

Feedback is particularly useful for behavior

that the learner can change.

The learning object strengthens visual

relationships between concepts and beliefs.

Concluding points


We assume that some of these principles may be more prominent than others when it comes

to develop persuasive learning designs, however, they are all considered important. Learners

emphasized the effects of a flexible design approach and identified a need to encourage

teachers and designers to use new media and differentiated learning scenarios for their

teaching. It is particularly interesting, however, that the overall types of activities that the

different pilot projects realised were quite similar, not least due to the same technologies

(PLOTMaker and PLOTLearner) that were used for their development.

It can be assumed that the different sets of learning designs and/or learning objects that they

had tested have been highly valuable for their learning and motivational throughout their

learning experience. The strategy of learning designers for the future should be to allow for

sufficient time during the learning experience to achieve persuasion.

Acknowledgements The EuroPLOT project was funded by the Education, Audiovisual and Culture Executive

Agency (EACEA) of the European Commission through the Lifelong Learning Program with

grant #511633.

References

Biggs, J.B. and Collis, K.F. (1982) Evaluating the quality of learning: the SOLO Taxonomy. Academic Press:

New York.

Boyle T. (2008). The design and development of learning objects for pedagogical impact. In Lockyer, l.,

Bennett, S. Agostinho S. and Harper B. (Eds) The Handbook of Research on Learning Design and Learning

Objects: issues, applications and technologies.

EuroPLOT (2010). Persuasive Learning Objects and Technologies. http://www.eplot.eu/

Fogg, B. J (2003). Persuasive Technology: Using Computers to Change What We Think and Do. Morgan Kaufmann Publishers.

Krauss, F., & Ally, M. (2005). A study of the design and evaluation of a learning object and implications for

content development. Interdisciplinary Journal of Knowledge and Learning Objects, 1 1-22.


Part B: Beyond EuroPLOT


Persuasive Design

The concept of persuasive design has many aspects to be considered. Four research papers address

some of these aspects. Gram-Hansen asks the question if persuasive design is a matter of context

adaptation. Torning has investigated the persuasive design of learning experiences. Grund-Sørensen

has developed a multi-modal interface for cultural exploration, and Wiafe has postulated a unified

framework for analysis, design and evaluation of persuasive technologies.


Persuasive Design – A Matter of Context Adaptation?

Sandra Burri Gram-Hansen Department of Communication and Psychology

Aalborg University

[email protected]

Abstract: Within the field of Persuasive Technology it is widely acknowledged that

successful persuasion is dependent on timing and the ability to act within the opportune

moment known as Kairos. Kairos constitutes the link between the opportune moment and the

appropriate action, all in consideration of the specific context. This paper argues that the

claim of persuasive design may not be the ability to change the user’s attitude or behaviour

towards a given subject, but rather on the ability to create designs, which adapt the context in

a way, which facilitates the ability to act within the opportune moment.

Keywords: Persuasive Design, Kairos, Context, Context Adaptation, Context

Awareness

Introduction

Based on findings within and beyond the EU funded e-PLOT project, this paper argues

towards a wider and more nuanced understanding of Persuasive Design (PD). Methodological

and theoretical research within the field of Persuasive Technology (PT) has so far primarily

focused on a further development and deeper understanding of a list of design principles

argued to hold persuasive potential(Fogg, 1998). However, once applied within more

established research fields such as Technology Enhanced Learning (TEL), Information

Architecture (IA) and Digital Mediation of Cultural Heritage (DMCH), these design

principles appear to be already widely applied. This lack of originality challenges the unique

claim of PD, and gives reason to look beyond specific design principles in order to clarify the

potential of PD in relation to existing HCI areas. This paper suggests that the claim of PD

may more reasonably be based on a wider understanding of the rhetorical notion of Kairos

(Hansen, 2009), and more specifically the acknowledgement that in order for any technology

to hold persuasive potential, there must be an appropriate and adaptive balance between the

technology and the context in which is applied.

As our acceptance of technology has increased, so has the influencing potential of these

devices. Amongst the newly established perspectives on technology application, which are

being explored is the notion of applying interactive computer technologies when intentionally

changing the attitude or behaviour of users. Most often, this approach to technology is

referred to as persuasive or motivational design. The definition of Persuasive Technologies

includes all types of interactive computer technologies designed with the intent to change

attitudes and behaviour, however location based and context aware technologies are credited

as holding a particular persuasive potential.


PD is most often described in continuance of work originally presented by Stanford

University researcher, BJ Fogg, who in 1998 introduced the notion of PT, and in 2003

published the first book on the subject (Fogg, 1998, 2003). Approaching the notion of

computers as persuaders from a social psychology perspective, Fogg suggests that computers

hold a particularly strong potential to change the attitudes and behaviours of the users, but

emphasizes that the designer cannot rely on coercion or deception when influencing the user.

As such, Fogg defines a specific perspective on interactive computer technologies, which

distinguishes itself from e.g. marketing technologies or technologies that somehow mislead

the users. By this definition, Fogg not only highlights an important link between PT and

classical rhetoric, he also emphasizes that ethical reflections must play a central role in

relation to this particular type of interactive technology.

Persuasive Design – Potential and limitations

The acknowledgement that designers are able to greatly influence the way a given situation is

perceived by the user, has encouraged researchers to explore and develop theoretical,

methodological and practical aspects of applying computers as persuaders. In particular, much

attention has been directed towards one aspect of Fogg’s research, known as The Functional

Triad. The triad constitutes a categorized framework in which Fogg has identified three

different psychological roles in which a persuasive technology may function, namely:

Tool (Technologies which ease the intended behaviour)

Media (Technologies which enable users to explore and experience the consequences

of given actions)

Social Actor (Technologies which provide social support)

Fogg refers to the functional triad as a framework for evaluating and understanding the user’s

experience of applying a technology, and it is suggested that the reflections regarding the

psychological role of the technology may be useful to researchers as they explore the notion

of persuasive technologies further. Furthermore, the functional triad is argued to be of value

to designers, as an inspiration to those who aim to develop persuasive technologies.

For each of the roles in the triad, Fogg lists a number of persuasive principles, which –

through analysis of a large variety of persuasive technologies - he identifies as system design

commonalities. In spite that these principles have been identified and categorized through

careful exploration of already existing technologies (and as such do not constitute novel

design solutions), they have become one of the primary points of interest for researchers who

have sought to further develop Fogg’s work.

Most renowned within the PT research community, is the Persuasive System Design model

(PSD), which was introduced by Oinas-Kukkonen and Harjumaa in the acknowledgement that

the functional triad lacked a design oriented perspective. The PSD model presents a

categorization of the persuasive principles of the functional triad, which establishes a link

between these principles and well-known features of requirement specifications, thereby

making the categories more apprehensible for system developers. (Oinas-Kukkonen, 2008).

Other principle based research approaches include, mapping the principles to areas such as

classical rhetoric and information architecture (Lykke, 2009; Pertou & Iversen, 2009), and


investigating and formulating individual principles and triad related perspectives, such as

credibility, praise and rewards (Bertelsen & Schärfe, 2010) (Anne Gerdes and Øhrstrøm,

2011).

Each of the theoretical and practical approaches to the further development of the PT

framework constitutes important perspectives in relation to establishing a deeper

understanding of the potential of PT. However, the focus on formalising the PD principles

also comprises a significant challenge once these theories and methods are applied within

other research areas. As the design principles have originally been identified through analysis

of already existing technologies, they themselves do not contribute with new approaches to

design. As a result, the unique claim and novelty of PT is questioned when the framework is

applied within more established research areas.

Persuasive or simply enhanced?

Several researchers have addressed the challenges related to Fogg’s original work, since the

establishment of the PT research field. One of the first to point towards aspects which might

be questionable, was Bernadine Atkins, who in 2006 published the paper “Captology: A

Critical Review”, in which she places a strong focus upon the ethics of PT and also argues

towards a distinction between education, advocacy and persuasion (Atkinson, 2006).

Whilst Atkins approach may have motivated others to place a stronger focus on the ethics of

PD, it can be argued that the PSD model exemplifies an attempt to not only develop and

applicable method for persuasive system design, but to also meet some of the implied

problems related to the applicability of Fogg’s original framework. However, while the PSD

model presents a more formalized understanding of the content of the functional triad, the

focus remains on the individual principles, rather than on the different roles of the triad, and

does as such differentiate itself from Fogg’s original recommendations regarding the triad.

In the process of exploring the cross field between PD and IA, Marianne Lykke comes to the

conclusion that whilst the persuasive principles can be related to IA components, the

principles themselves do not lead to any extraordinary design ideas (Lykke, 2009). Thereby

Lykke’s conclusion exemplifies one of the fundamental challenges when applying the existing

theoretical and methodological framework for PT to more established fields; namely that the

framework does not provide argumentation towards a unique claim for PD, but merely

comprises a reflective meta-layer to existing design approaches.

Similar conclusions were reached within the EU funded e-PLOT project, as steps were taken

towards defining a theoretical and practical definition of persuasive learning designs. Several

researchers, including Fogg and Atkins, have suggested that an overlap exists between

learning and persuasion, but one of the challenges within the e- PLOT project has been to

argue towards areas in which PD may in fact enhance existing learning technologies to an

extent where the technologies may be defined as persuasive. The theoretical framework

sought to define the cross field between the existing PD framework and a constructivist

approach to learning, and in practice the two technologies which were included in the project

were analysed from a PD perspective in order to determine areas in which PD might


contribute to the learning process. The acknowledgement that learning is something which

takes place within an individual, and that we construct knowledge based on our

experiences(Biggs & Tang, 2007), shares a number of commonalities with the notion of

persuasion, where endogenous motivation is regarded as particularly important(Fogg, 2003).

It became evident that both in theory and practice, the persuasive principles were already

widely at use in technology enhanced learning, and as such offered no new perspectives to the

further development of the technologies. From a theoretical perspective the approach to PD

sprung from Fogg’s original work, but was supported by perspectives from classical rhetoric

and temporal logic and more nuanced definition of persuasion (S. B. Gram-Hansen, 2012).

Subsequently the approach to PD was related to a constructivist approach to learning, in the

acknowledgement that this particular pedagogical approach showed several similarities to the

theoretical framework for PD (S. B. Gram-Hansen, 2012).

Likewise, on-going research in the cross field between PD and digital mediation of cultural

heritage argue that the persuasive principles do not constitute a unique claim for PD, nor does

the notion of changing the attitudes or behaviours of the users. Cultural heritage can be seen

as identity-forming and –reproducing and aiming to re-establish the lost relation to our past

and rediscover our roots (Lund, 2004; Lund, Andersen, Christensen, Skouvig, & Johannsen,

2009). E.g. the former Danish Ministry of Cultural Heritage was created in order to rediscover

the cultural roots of the Danes in a project that can be viewed as nationalistic in its aim: to

establish a national identity. As such, the project exemplifies that the concept of attitude

change is also an intended outcome which is focused upon in areas beyond the field of

Persuasive Technology and PD.(S. B. a. G.-H. Gram-Hansen, Lasse Burri, 2013)

With several researchers from different areas coming to the same conclusions about the

limitations of the primary elements in the existing PD framework, it appears reasonably to

take a step back and work towards an understanding and potential definition of PD, which

acknowledges and incorporates the work already achieved within the PT field, but which also

clarifies the areas in which PD may potentially enhance the design and development of

interactive computer technologies in different application areas.

A matter of appropriate balance

In exploration of the notion of persuasion, and in the aim of extending the theoretical

foundation of Persuasive Designs, some researchers approach the challenges of this novel

field from a foundation in classical humanistic traditions such as rhetoric, logic and ethics.

The very idea of persuasion is commonly considered as having been brought into the world

by classical rhetoric. Classical rhetoric has been systematically related to social psychology

by Michael Billig (Billig, 1996). A central statement in Billig’s Arguing and Thinking, is that

we may gain significant insight into human perception by exploring argumentation and

especially by studying what classical rhetoric can contribute to the field. Billig observed that

social psychology had had a tendency to identify thinking with rule following. From classical

rhetoric he learned, however, that while arguments and thought may well be based on rules,

rules themselves arise from arguments, and indeed, may be disputed by arguments, that is to

say that while rules do exist, they are not deterministic. One should not rely on the assumption


that following certain rules will always yield the desired results (Hasle, 2007; Pertou &

Iversen, 2009).

One of the concepts, which are generally accepted as being a key element and a requisite for

successful persuasion, is the rhetorical notion of Kairos. Kairos is often described as timing,

or the ability to perform the appropriate action at the right time and in the right place. In term

of appropriate, the performed action is required to be not only effective but also ethical. The

concept sums up the principle that any rhetorical approach is based upon the specific

situation, and that comprehension of the context as such is one of the most vital resources

when deciding upon rhetorical means to apply to a given argument (Hansen, 2009). Hansen

specifies that the definitions of Kairos vary from narrow translations such as “particular point

in time” and “specific circumstance”, to wider concepts such as “situation”, “occasion” and

“opportunity”.

When relating the different meanings of Kairos in a PD context, the narrow definition serves

well in relation to specific design related choices, such as determining the appropriate time for

initiating a persuasive strategy (i.e. triggering a specific behaviour), an argument which has

been raised by several researchers over the years (Aagaard, 2008; Glud & Jespersen, 2008)

The wider definition on the other hand, supports the argument that in order to successfully

select and apply a persuasive principle to the design of a technological device, the designer

must beforehand acquire a fundamental understanding of the context in which the device is to

be applied, and use this knowledge to create a technology which will be appropriate to the

given situation.

Kairos in itself is a powerful and multifaceted concept that is not easily formalized. As such,

even though Kairos is vital in relation to successful persuasion, the concept in itself does not

translate easily to the digital context of persuasive technologies. However, the challenges

related to integrating the notion of Kairos in the development of interactive technologies, may

be addressed by considering Arthur Priors perspectives on temporal logic. More specifically,

the development of PDs may benefit greatly from Priors arguments that time is not only a

specific moment but also a wider contextual concept, which he distinguished between as A-

time and B-time Priors notion of B-time refers to the objective perception of time, which has

dominated the philosophical and the scientific debate for centuries and which is expressed by

for instance traditional calendars. A-time on the other hand refers to the contextual perception

of the present moment, and takes into consideration the imbalances which are caused by

previous events (Øhrstrøm, 1995).

Kairos as it is described by Hansen (2009), may be related to A and B time, by considering

Priors notion of A-time as the formalization of Hansen’s wider definition of Kairos, whilst

Priors notion of B-time may be related to Hansen’s narrow definition of the concept. Glud and

Jespersen elaborates upon the importance of considering not only Kairos but also Prior’s

notion of A- and B-time in the development of persuasive systems, in a conceptual analysis of

Kairos in relation to location based services. They conclude that inclusion of Kairos in the

development of mobile persuasive technologies is spatiotemporal and demand that all

conceivable time dimensions are taken into consideration (Glud & Jespersen, 2008)


The different perspectives of Kairos presented by Hansen (2009) are inseparable in the respect

that both must be taken into consideration when determining the appropriate moment to

initiate a persuasive action. Likewise A- and B- time cannot be considered separate, but must

both be taken into account when designing persuasive systems. In order to fully conceive the

notion of appropriate timing, one must include both a broader understanding of the defined

aim of the process, and consider the contextual reality of the user whilst the steps of the

system is being completed.

It may be argued that a similar correlation or balance exists in relation to the concept of

intentionality. PT is originally defined as interactive technologies that are designed with the

specific intention to change the user’s attitude and/or behaviour (Fogg, 2003) and as such the

intention itself becomes a central element in the understanding of PD. Designers will most

often have a specific and quite complex intention with the design of a technology. This

intention must to some extent be balanced towards the user’s intention behind applying the

technology – an intention that is most likely influenced by exogenous factors. In consideration

of the designer most often having a specific use context in mind when developing a

technology, it may reasonably be argued that both context and design of the specific

technology must be taken into consideration if such technology is to facilitate a persuasive

purpose.

The notion of approaching the concept of PD as being multi-layered has furthermore been

included in Fogg’s original framework. When introducing the concept of persuasive

technologies, Fogg distinguished between Macrosuasion and Microsuasion as a way to

explain and clarify the dynamics of persuasive technologies. The distinction between the two

is important in terms of both analysis and development of PD in most computer technologies.

The term Macrosuasion describes an overall persuasive intent of a technology, whilst

Microsuasion refers to the use of PD principles in technologies which do not necessarily have

en overall persuasive goal (Fogg, 2003). As was the case with temporal logic, Macrosuasion

and Microsuasion can reasonably be related to the definition of Kairos as defined by Jette

Hansen. The primary distinction between Fogg’s definition of Microsuasion and

Microsuasion, in relation to the presented definition of Kairos and the reflections concerning

the persuasive intention, is that whilst Fogg argues that Microsuasion may be applied in

technologies which holds no macrosuasive intention, the latter two perspectives insist that

both the wider and the narrow perspective must be considered if the persuasive intention is to

be fulfilled.

Persuasive Design – A Matter of Context Adaptation

In the previous section, it is exemplified how the multi-layered perception of Kairos, may be

linked to several of the perspectives which have been addressed as relevant to the further

development of the PT and PD research field. Individually, these different examples do not

provide sufficient insight to clearly define a unique claim of PD, however, the commonality

between the different perspectives does motive a further consideration of the notion of

ensuring an appropriate balance.


As mentioned, Kairos is widely acknowledged as a requisite for persuasion, and as such, a

successful persuasive design must necessarily strive to comply with all three dimensions of

the concept. In the process of doing so, location based and context aware systems have been

credited as holding a particular persuasive potential, as they enable the design to incorporate

location and time in the design. In 2007 mobile phones were argued to be the potentially most

efficient persuasive technology, as they not only hold a particular technological potential, they

also benefit from a unique affection from their users and have become the device which is

brought into almost any situation, and widely accepted by others. (Bødker, 2012; Fogg &

Eckles, 2007).

As such, various technological devices can be argued to hold the potential to meet both the

timely and location based dimension of Kairos, however the persuasiveness of each

technology is dependent on the device also being applied in the appropriate manner within the

given context. This third and final dimension calls for a different and more nuanced

evaluation of the intended use context, it may not be formalized and it may be argued to point

towards the necessity of not only designing the appropriate technology, but also design the

appropriate balance between the technology and the intended use context.

Previously it has been argued that the intentionality of the designer is not only directed

towards the technology, but also towards the intended use context, and that by adding a

technology to a given context, the user’s perception of that context is altered (S. B. Gram-

Hansen, 2012). In consideration of the notion of appropriate manner, it may be that this

perspective should be extended, in the sense that PD not only acknowledges the impact that

the technology has on the context, it aims to create the appropriate balance between

technology and context.

By doing so, PD may be considered an approach to design that particularly incorporates the

notion of Context Adaptation (CA). CA is most often referred to in relation to development of

context and location aware systems, as these devices actively adapt information about the

contexts in which they are applied. However, considering the impact that a technology may

have on a given context, CA in relation to PD may be considered a reciprocal balance, were

the technologies adapt the appropriate information about the context, and the context is

designed to adapt the technology.

In practice, this approach to PD distinguishes itself from PT, in a way that acknowledges the

theoretical and practical steps taken so far within the PT research community, by

distinguishing between PT and PD approaches. System oriented methods such as the PSD

model address the challenges related to the specific technology design, where as PD is

considered a wider concept which focuses on the establishment of an appropriate balance

between technology and context, and which may serve as a meta-perspective to more

established research fields.

References

Aagaard, Morten and Peter, Øhrstrøm and Lars, Moltsen. (2008). It might be Kairos. Paper

presented at the Persuasive 08, Oulu Finland.


Anne Gerdes and Øhrstrøm, Peter. (2011). The Role of Credibility In the Design of Mobile

Solutions To Enhance the Social Skill-Set of Teenagers Diagnosed with Autism.

Journal of Information, Communication and Ethics in Society, 9(4), 12.

Atkinson, Bernadine M. C. (2006). Captology: A Critical review. Paper presented at the

Persuasive Technology 2006.

Bertelsen, Lykke Brogaard, & Schärfe, Henrik. (2010). Tracing concepts in Designing for

Change. Paper presented at the The 14th World Multi Conference on Systemics,

Cybernetics and Informatics, Proceedings Volume III International Institute of

Informatics and Systemics (IIIS), Orlando, Florida.

Biggs, John, & Tang, Catherine. (2007). Teaching for Quality Learning at University (3rd

ed.): Society for Research into Higher Education & Open University Press.

Billig, Michael. (1996). Arguing and Thinking. A Rhetorical approach to social psychology:

Cambridge University Press.

Bødker, Susanne and Ellen Christiansen. (2012). Poetry in motion – Appropriation of the

world of Apps. Paper presented at the The 30th European Conference on Cognitive

Ergonomics, ECCE '12.

Fogg, BJ. (1998). Persuasive Computers, Perspectives and research directions. Paper

presented at the CHI.

Fogg, BJ. (2003). Persuasive Technology, Using Computers to change what we Think and

Do: Morgan Kaufmann Publishers.

Fogg, BJ, & Eckles, Dean. (2007). Mobile Persuasion - 20 Perspectives of the Future of

Behavior Change: Persuasive Technology Lab, Stanford University.

Glud, Louise Nørgaard, & Jespersen, Julie, Leth (2008). Conceptual analysis of Kairos for

Location-based mobile devices. University of Oulu. Department of Information

Processing Science. Series A, Research Papers, 17-21.

Gram-Hansen, Sandra Burri. (2012). Deliverables Report D.3.3, Persuasive Learning

Designs: Aalborg university.

Gram-Hansen, Sandra Burri and Gram-Hansen, Lasse Burri. (2013). On the role of ethics in

Persuasive Design. Paper presented at the Ethicomp 2013, Kolding.

Hansen, Jette Barnholdt. (2009). Den rette tale på det rette tidspunkt. RetorikMagasinet, 74.

Hasle, per and Anne-Katrine Kjær Christensen. (2007). Classical Rhetoric and a Limit to

Persuasion. Paper presented at the Persuasive Technology, Palo Alto.

Lund, Niels D. (2004). Kulturarv - et begreb mellem politik og information.

Biblioteksarbejde, 67/68.

Lund, Niels D., Andersen, Jack, Christensen, Hans Dam, Skouvig, Laura, & Johannsen, Carl

Gustav. (2009). Digital formidling af kulturarv: Fra samling til sampling. København,

DK: Forlaget Multivers.

Lykke, Marianne. (2009). Persuasive design strategies: means to improve the use of

information organisation and search features in web site information architecture?

Paper presented at the Proceedings ASIST Special Interest Group on Classification

Research 20th Workshop, Vancouver.

Øhrstrøm, Peter and Per, Hasle. (1995). Temporal Logic: From Ancient Ideas to Artificial

Intelligence (Vol. 57): Kluwer Academic Publichers.

Oinas-Kukkonen, Harri and Marja Harjumaa. (2008). A systematic Framework for Designing

and Evaluating Persuasive Systems. Paper presented at the Persuasive 2008, Finland.

Pertou, Maria, & Iversen, Signe Dahl. (2009). Persuasive Design i Retorisk Perspektiv.

rhetorica Scandinavia, 46/50, 15.


Persuasive Design of TEL: Models and Challenges

Kristian Torning Danish School of Media and Journalism, Emdrupvej 72, DK-2400 Copenhagen

[email protected]

This paper reviews the emerging research literature on persuasive design models, while at the same time counterbalancing it with a comparative review of the most popular models as well as findings from the field of Technology Enhanced Learning (TEL). These models are systematically compared and analysed with regard to producing technology enhanced learning experiences. The results of this comparison can be used when selecting models for designs, in addition to concurrently serving as an offset for creating new models.

Keywords: Persuasive Technology Design, TEL, Design Models, Persuasion

Introduction

Persuasive design (PD) is centred on the concept that humans may delegate the act of persuasion to non-human entities, deliberately seeking to change or reinforce behaviours and attitudes. In 400 B.C., Aristotle defined rhetoric as “…the faculty of observing in any given case the available means of persuasion” (Aristotle 2010). PD (Redström 2006) is increasingly becoming an area of interest across multiple domains. Persuasion can be used in multiple domains, and in observing how Aristotle, Cicero and other modern scholars (Dillard, Pfau 2002) describe persuasion, it soon becomes evident that ‘persuasion’ is not the automatic result of certain steps engaged in by an individual. On the contrary, persuasion is a complicated and somewhat unpredictable social process tied to the rhetorical situation in which persuaders seek to persuade (Bitzer 1968, Vatz 1973, Torning 2008). In the field of PD, the topic of ‘health’ has especially attracted particular attention (Torning, Oinas-Kukkonen 2009, Chatterjee, Price 2009, Kamal, Fels 2012). Problems such as alcoholism, obesity and lack of exercise all provide a multitude of interesting research questions, since their singular real remedy lies in individuals changing their behaviour patterns and remaining consistently motivated towards maintaining the behavioural change thus initiated (WHO Regional Committee for Europe 2008, IJsselsteijn et al. 2006). As with ‘health’, ‘learning’ is also a complicated social process dependent on the behaviour and attitude change of individuals, as well as continuous self-motivation. However learning may not occur even if learners are both motivated and have a positive attitude towards learning. We can speculate that addressing ‘learning’ with persuasive design in numerous settings might be a more difficult proposition than addressing ‘health’. Like persuasion, learning too is rooted in ancient Greece and rhetoric. Subsequently, Cicero went on to describe the three functions of the rhetorician: “Docere, Delectare, Movere,” or “To Teach, To Delight, To Move.” (Christensen, Hasle 2007). Thus, it becomes obvious that teaching has been an integrated part of rhetoric for many years. Persuasion and learning are very situational and dependent on the interplay of many different factors (Wang, Hannafin 2005). For this very same reason, we cannot expect to find a unified theory or model of either concept, as they are bound by a certain context comprised of (for example): student/audience, teacher/speaker, classroom/venue, topic, values, zeitgeist and culture. In the domain of learning, a relatively recent change is the adoption of interactive technologies (especially Learning Management Systems, which are being increasingly adopted by various institutions). Wang and Hannafin (2005) offer a broad definition of such technologies: “Technology Enhanced Learning


Environments (TELEs) are technology-based learning and instructional systems through which students acquire skills or knowledge, usually with the help of teachers or facilitators, learning support tools, and technological resources.” In combining PD with TEL, we are able to identify the outline of a problem that is characterised by a staggering level of complexity. How can we begin to address these many different variables combined with nuances of numerous abstract social concepts? What exactly are practitioners and researchers expected to accomplish, if they wish to design a persuasive TELE? Typically, researchers have used PD to address several health educational issues. Some examples of PD would be a computerised doll designed to teach adolescents about the consequences of not using birth control (Realityworks 2010, Fogg 2003); teaching women in rural India about their menstrual cycle and personal hygiene (Parmar, Keyson & deBont 2008); changing office workers' sitting habits in order to prevent back strain (Obermair et al. 2008) and designing kitchens that promote calorie-awareness cooking (Chi et al. 2008). It is interesting to note that all these studies were conducted by researchers who were not specialists in the field of education. If researchers are to address ‘learning’ and education at the macro level of persuasion (Fogg, 2003), they should design models that difficult as this may be, specifically address the rhetorical situation surrounding ‘learning’ and include findings from educational research including TEL.

Methodology

If researchers aim at developing a TEL specific PD-model tailored to address the domain of learning and technology, an accepted starting point is a thorough review of some existing PD-models. In this paper, PD-models are coupled with TEL, by subjecting PD-models to evaluation employing the nine TEL research design principles developed by Wang and Hannafin (2005). The authors arrive at the nine principles as a result of their literature review of design-based research paradigms and state that: “To generate practical, credible, and contextual design theories, however, rigorous, disciplined, and iterative inquiry is needed […] we identify nine principles central to planning and implementing TELE design-based research” (p.15). In this paper however, the principles are applied to evaluate design models and in that sense one could argue, that the principles are being stretched beyond their intended domain. However, it is debatable if design models can be regarded as a form of meta-designs, and if they are in addition, expected to comply with sound standards of design research. The analysis presented here can be seen as a small first step towards creating a TEL centric PD-model and the resultant overview of PD-models is also intended to make the task of selecting a design strategy easier, for those wishing to mix persuasion, learning and technology.

Persuasive Design Models

Since PD is a fledgling area of research, one does not have the option to choose from numerous existing models. Three frameworks were selected for the purpose of this analysis: The Design with Intent (DwI) Method (Lockton, Harrison & Stanton 2010); the Persuasive Systems Design (PSD) Process Model (Oinas-Kukkonen, Harjumaa 2009); and the Eight-Step Design Process (Fogg 2009). These models can be considered first generation PD-models and they were selected based on general work in the field of PD. The three models originate from the Persuasive Conference Series, which is currently the main outlet for scientific dissemination centred on PD. Emerging PD-models created by practitioners (Futerra 2013) or a mix of practitioners and researchers (Fabrique 2013) were omitted. Although Fogg’s first model: The Functional Triad (Fogg 2003) has often been widely used, it was omitted in favour of his newer 2009 design process.


Design with Intent (DwI) Method Early on in the Persuasive Conference series Dan Lockton et al. addressed behaviour change from a perspective dubbed “Design with Intent,” (DwI) defined as “design intended to influence or result in certain user behaviour” (Lockton, Harrison & Stanton 2008, Lockton et al. 2009, Lockton 2013). The DwI Method is intended to be generally applicable in influencing user behaviour. The latest iteration of the model (Lockton, Harrison & Stanton 2010) is comprised of two modes: ‘Inspiration’ and ‘Prescription’.

In the 'Inspiration' mode, the designer takes inspiration from a set of headline design patterns, which are applicable to a wide range of target behaviours, grouped into six different ‘lenses’, representing particular disciplinary perspectives on using design to influence behaviour. In the ‘Prescription’ mode, the designer formulates a range of target behaviours (intended outcomes), describing interactions and as a consequence, a subset of the most applicable design patterns from each ‘lens’ is presented for each target behaviour,. The total number of patterns will vary depending on the chosen target behaviour(s) (according to the authors' typically 15–25 applicable patterns). This mode effectively ‘prescribes’ a set of patterns, which are deemed especially applicable or have already been applied to similar problems by other designers, in other contexts.

The ‘lenses’ themselves, are a way of grouping design patterns which share similar considerations, behavioural understanding or assumptions about how to influence users: to some extent, these groups resolve into particular ‘worldviews’, the way that a designer versed in a particular discipline might approach a brief on influencing behaviour. The six ‘lenses’ thus target specific domains: 1) The “Architectural Lens” is based on techniques utilised to exert influence on user behaviour in architecture, urban planning and related disciplines. 2) The “Errorproofing Lens” tackles deviations from the target behaviour by treating them as ‘errors,’ which design facilitates in evading, either by making it more uncomplicated for users to work, with no possibility of being prone to errors. 3) The “Persuasive Lens” makes use of computers with interfaces in order to convince users into transforming attitudes and behaviour. 4) The “Visual Lens” seeks to merge ideas from product semantics, semiotics, ecological psychology and Gestalt psychology on the subject of the manner by which users distinguish patterns and meanings. 5) The “Cognitive Lens” is based on research in behavioural economics and seeks to understand by what method people make decisions. 6) The “Security Lens” symbolises a ‘security’ worldview, which states that it is possible to discourage and/or avert unacceptable user behaviour. When designers seek to employ the ‘lenses,’ a free cards based solution is available. The cards serve as a very tangible way to explore the various design patterns (Lockton 2013).

Persuasive Systems Design (PSD) Model Oinas-Kukkonen & Harjumaa (Oinas-Kukkonen, Harjumaa 2009) offer a model grounded in previous research. The model is comprised of three distinct phases: 1) Understanding key issues behind persuasive systems, 2) Analysing the persuasion context and 3) [Selecting] Design of system qualities. The optimal result of this linear model is “Behaviour and/or attitude change”.

The first phase in the words of the authors addresses the understanding of basic concerns fundamental to special persuasive information technology sytems, prior to implementing these systems. Oinas-Kukkonen & Harjumaa (Oinas-Kukkonen, Harjumaa 2009) state that, “Only after obtaining a reasonable level of this understanding can the system be analysed and designed.” Seven postulates behind persuasive systems are also offered.


In the second phase, seven core elements are offered, to create a better understanding of the context of persuasion:

1. Persuader: the designer who is deliberately seeking to change the behaviour or attitude of the system user.

2. Change type: the type of change in behaviour and/or attitude the designers aim at invoking via their design.

3. Use context: the features arising from the problem domain target, which the persuasion design addresses e.g. features specific to ‘health’ or ‘learning’.

4. User context: the traits of the targeted user, e.g. goals (including current progress toward achieving them, and potential past performances), commitment, lifestyle etc.

5. Technology context: The strengths and weaknesses, as well as the risks and opportunities, of specific technological platforms, applications and features.

6. Message: the form and content delivered to the user who has to be persuaded. The form is how the message is presented, e.g. as raw text, in a dialogue, or in a game. The content of the message has to fit the form.

7. Route: persuasion can be direct, indirect or both. A direct route would be one wherein the message contains only a few strong arguments, while an indirect route has numerous arguments.

In the third phase of the design model, the qualities of the information system being designed are modelled. Here the designer can select from an extensive catalogue listing 28 design principles for persuasive system content and functionality. The principles are ordered in the following main categories: primary task support, dialogue support, system credibility support and social support. For each category, several approaches are offered with concrete advice and examples of implementation.

The Eight-Step Design Process Benjamin J. Fogg can be regarded as one of the founding fathers of ‘persuasive technology’ and at least, the most widely cited (Fogg 1998, Fogg 1999, Fogg 2003). In his eight step model, he states that, “The goal of Steps 1 through 7 of the design process is to create a digital product that reliably persuades someone — not everyone — to adopt the target behaviour.” Following the steps, the designer (or design team) is guided towards creating a persuasive technology. However, Fogg asserts that, “The eight steps are not intended to be a rigid formula; instead, the steps serve as milestones to make the design process more effective.”

1. Choose a simple behaviour to target: Here designers decide on an appropriate behaviour to target for change.

2. Choose a receptive audience: If the project does not demand otherwise, it is best to ensure that the audience is familiar with the technology channel, thus making them more responsive to change.

3. Find out what is preventing the audience from performing the target behaviour: As reported by Fogg, it may be either a) lack of motivation, b) lack of ability or c) lack of a well-timed trigger to perform the behaviour. However, the design process offers little evidence, which makes it difficult to assess that claim.

4. Choose a familiar technology channel: The “best” channel usually depends on three factors: the target behaviour, the audience, and what is preventing the audience from adopting the behaviour—i.e., the first three steps in the design process


Optional (unnumbered) step: Re-ordering the First Four Steps: Designers usually perform the first four steps in sequence, but according to Fogg, in some cases designers will make an exception and carry out the steps in a different order.

5. Find suitable examples of persuasive technology: Search for examples of successful persuasive technologies that are relevant to the intervention, by seeking out solutions that succeed in getting people to change their behaviours.

6. Imitate successful examples: Rather than going back to square one, a more appropriate and desirable approach is to imitate successful examples of what is already working.

7. Test and iterate quickly: After imitating successful examples, test the designed user experience promptly and repetitively. A series of small, very quick tests (not scientific experiments) is preferable to one big test. The purpose is rapid prototyping, best achieved by measuring behaviour.

8. Expand on success: Even a small behaviour change is a milestone. The technique of expansion should be systematic, by varying only one or two attributes from the success achieved in Step 7. Fogg states that as perceived from a scientific perspective, step 8 is the starting point for a controlled experiment.

Results and recommendations

The application of Wang and Hannafin's nine principles for design based TEL research served to shed light on various traits of the PD models. Each model was evaluated against the nine principles to reveal their strengths and weaknesses as outlined in Table 1 below. Some common issues were discovered in all the models. For instance, with regard to Principle 1 (“Support Design with Research from the Outset”), it is surprising that all of the models largely ignore the opportunity to integrate previous findings from the research areas of communications, rhetoric and ethics, as these areas predate research on PD and should be specifically addressed when presenting any model prescribing how to conduct ‘persuasion’. The models also presume that anyone can generate data and analyse the user context for persuasion. Principle 5 (“Implement Research Methods Systematically and Purposefully”) and Principle 6 (“Analyze Data Immediately, Continuously, and Retrospectively”) are not addressed by any of the models, which thus display a common weakness, by omitting to explicitly concentrate on and tackle the generation and analysis of empirical data for informing a persuasive design. Perhaps the authors take these activates for granted. When informing a design from the User Centred-Design (Beyer, Holtzblatt 1998, Holtzblatt, Wendell & Wood 2005) prospective researchers are dominantly seeking to support tasks' completion by lowering the cognitive burden of those tasks, whereas PD seeks to transform and motivate behaviours. It appears counterintuitive that such an apparently large difference and distinction in objectives, should not somehow affect how researchers generate data and analyse it when informing the designs. However, developing clear PD research guidelines for data generation with regard to observations, interviews, surveys, document analysis or field studies etc. is a daunting task, and perhaps it would be premature, if researchers were to address that issue before designing more robust models. The Design with Intent (DwI) model and the Persuasive Systems Design (PSD) model both fail to address Principle 7: (“Refine Designs Continually”), and therefore, testing and social interactions with users are not addressed. Another general issue with the models is that ‘persuasion’, supposedly resides in a right mix of strategies for the right rhetorical situation by creating a fitting (technology) response (Bitzer 1968, Torning 2008). When we speak of


Tabl

e 1.

Per

suas

ive

Des

ign

fram

ewor

ks m

easu

red

agai

nst W

ang

and

Han

nifin

's n

ine

desi

gn p

rinc

ples

.

Prin

cipl

es 1

-9

(Wan

g, H

anna

fin 2

005)

Des

ign

with

Int

ent (

Dw

I) M

etho

d (L

ockt

on, H

arris

on &

Sta

nton

20

10)

Pers

uasi

ve S

yste

ms

Des

ign

(PSD

) M

odel

(O

inas

-Kuk

kone

n, H

arju

maa

200

9)

Eig

ht-S

tep

Des

ign

Proc

ess

(Fog

g 20

09)

1: S

uppo

rt D

esig

n w

ith

Res

earc

h fr

om th

e O

utse

t [I

nteg

ratin

g pr

evio

us fi

ndin

gs

i.e. b

y lit

erat

ure

revi

ew]

• C

lear

ly o

ffer

s m

any

refe

renc

es to

pr

evio

us w

ork

from

a p

leth

ora

of

rese

arch

fiel

ds. T

he m

odel

's pe

rspe

ctiv

e on

‘des

ign’

is m

ade

very

cl

ear.

• Ea

ch ‘l

ens’

is b

acke

d up

by

clea

r re

fere

nces

.

• O

ffer

s a

varie

ty o

f ref

eren

ces

anch

orin

g th

e w

ork

mai

nly

in In

form

atio

n Sc

ienc

es, P

sych

olog

y, S

ocia

l Ps

ycho

logy

and

Hum

an C

ompu

ter

Inte

ract

ion

(HC

I).

• B

uild

s m

ainl

y on

the

auth

or’s

per

sona

l ex

perie

nce.

O

ffer

s au

thor

's w

ebsi

te a

s a

refe

renc

e,

whe

re s

ever

al re

fere

nces

are

list

ed.

How

ever

it is

unc

lear

, how

thes

e re

fere

nces

wer

e in

tegr

ated

into

the

desi

gn

proc

ess.

2: S

et P

ract

ical

Goa

ls fo

r Th

eory

Dev

elop

men

t and

D

evel

op a

n In

itial

Pla

n

• D

oes

not e

xplic

itly

men

tion

or

addr

ess

theo

ry d

evel

opm

ent.

• O

ffer

s ve

ry c

lear

rese

arch

goa

ls fo

r th

e m

odel

by

offe

ring

a se

ctio

n (2

.1.3

) ela

bora

ting

on th

e ev

olut

ion

of th

e m

etho

d.

• Ex

plic

itly

addr

esse

s th

eory

de

velo

pmen

t i.e

. “Th

is a

rticl

e is

co

ncep

tual

and

theo

ry-c

reat

ing

by it

s na

ture

…”

• C

lear

ly s

ynth

esiz

es s

ever

al th

eorie

s (th

at a

re d

escr

ibed

as

such

). •

Off

ers

very

cle

ar re

sear

ch g

oals

with

re

gard

to d

evel

opin

g th

e m

odel

.

• D

oes

not e

xplic

itly

men

tion

or a

ddre

ss

theo

ry d

evel

opm

ent,

but r

athe

r ref

ers

to

the

desi

gn p

roce

ss a

s ‘b

est p

ract

ice’

: “…

in th

is p

aper

I dr

aw o

n m

y 15

yea

rs

of e

xper

ienc

e in

stu

dyin

g an

d cr

eatin

g pe

rsua

sive

tech

nolo

gies

to o

ffer

wha

t I

cons

ider

to b

e “b

est p

ract

ices

”…”

3: C

ondu

ct R

esea

rch

in

Rep

rese

ntat

ive

Rea

l-Wor

ld

Setti

ngs

• O

ffer

s a

com

para

tive

clea

r des

ign

case

usi

ng th

e de

sign

of a

n A

TM

mac

hine

. It i

s ho

wev

er a

n ar

mch

air

stud

y (n

o us

ers

invo

lved

). Th

e be

havi

our t

arge

ted

is to

ens

ure

that

pe

ople

don

’t fo

rget

thei

r cre

dit c

ard,

w

hen

with

draw

ing

mon

ey.

• O

ffer

s an

exa

mpl

e de

sign

(Nik

e 20

13)

in th

e fo

rm o

f a th

orou

gh a

nd

conv

inci

ng a

naly

sis

empl

oyin

g th

e Pe

rsua

sive

Sys

tem

s m

odel

.

• D

oes

not o

ffer

rese

arch

repr

esen

tativ

e of

a re

al-w

orld

set

ting,

but

use

s pl

ausi

ble

thou

ght u

p ex

ampl

es to

ex

empl

ify s

ome

poin

ts.

4: C

olla

bora

te C

lose

ly w

ith

Parti

cipa

nts

• Th

e M

etho

d w

as d

evel

oped

thro

ugh

a se

ries

of w

orks

hop

sess

ions

(with

de

sign

stu

dent

s an

d re

cent

gr

adua

tes)

. •

Doe

s no

t off

er c

lear

gui

delin

es fo

r in

volv

ing

user

s in

app

licat

ions

of t

he

Dw

I mod

el.

• Th

e Pe

rsua

sive

Sys

tem

s D

esig

n M

odel

w

as d

evel

oped

by

two

rese

arch

ers

and

does

not

men

tion

valid

atio

n w

ith

othe

rs.

• Th

e m

odel

doe

s no

t off

er c

lear

gu

idel

ines

for i

nvol

ving

use

rs in

ap

plic

atio

ns o

f the

mod

el.

• Th

e D

esig

n Pr

oces

s is

cen

tred

on th

e ac

tiviti

es o

f a ‘d

esig

n te

am’,

but w

as

deve

lope

d by

a s

ingl

e re

sear

cher

, by

draw

ing

upon

his

per

sona

l exp

erie

nces

an

d do

es n

ot m

entio

n va

lidat

ion

with

ot

her d

esig

ners

/pra

ctiti

oner

s.


Prin

cipl

es 1

-9

(Wan

g, H

anna

fin 2

005)

Des

ign

with

Int

ent (

Dw

I) M

etho

d (L

ockt

on, H

arris

on &

Sta

nton

20

10)

Pers

uasi

ve S

yste

ms

Des

ign

(PSD

) M

odel

(O

inas

-Kuk

kone

n, H

arju

maa

200

9)

Eig

ht-S

tep

Des

ign

Proc

ess

(Fog

g 20

09)

5: Im

plem

ent R

esea

rch

Met

hods

Sys

tem

atic

ally

and

Pu

rpos

eful

ly

• Th

e M

etho

d do

es n

ot o

ffer

adv

ice

or

rese

arch

gui

delin

es e

.g. o

bser

vatio

ns,

inte

rvie

ws,

sur

veys

, doc

umen

t an

alys

is o

r fie

ld s

tudi

es e

tc.

• Th

e Pe

rsua

sive

Sys

tem

s D

esig

n (P

SD)

Mod

el d

oes

not o

ffer

adv

ice

or re

sear

ch

guid

elin

es e

.g. o

bser

vatio

ns,

inte

rvie

ws,

sur

veys

, doc

umen

t ana

lysi

s or

fiel

d st

udie

s et

c.

•

• Th

e D

esig

n Pr

oces

s do

es n

ot o

ffer

ad

vice

or r

esea

rch

guid

elin

es e

.g.

obse

rvat

ions

, int

ervi

ews,

sur

veys

, do

cum

ent a

naly

sis

or fi

eld

stud

ies

etc.

6: A

naly

se D

ata

Imm

edia

tely

, C

ontin

uous

ly, a

nd

Ret

rosp

ectiv

ely

• Th

e M

etho

d do

es n

ot o

ffer

adv

ice

or

guid

elin

es fo

r dat

a an

alys

is.

• Th

e M

etho

d do

es n

ot o

ffer

adv

ice

as

to h

ow to

mea

sure

‘per

suas

iven

ess’

.

• Th

e Pe

rsua

sive

Sys

tem

s D

esig

n (P

SD

Mod

el d

oes

not o

ffer

adv

ice

or

guid

elin

es fo

r dat

a an

alys

is.

• Th

e Pe

rsua

sive

Sys

tem

s D

esig

n (P

SD

Mod

el d

oes

not o

ffer

adv

ice

as to

how

to

mea

sure

‘per

suas

iven

ess’

.

• Th

e D

esig

n Pr

oces

s do

es n

ot o

ffer

ad

vice

or g

uide

lines

for d

ata

anal

ysis

. •

The

Des

ign

Proc

ess

does

not

off

er

advi

ce a

s to

how

to m

easu

re

‘per

suas

iven

ess’

.

7: R

efin

e D

esig

ns C

ontin

ually

•

Doe

s no

t exp

licitl

y re

com

men

d th

at

desi

gner

s w

ork

in a

n ite

rativ

e fa

shio

n. T

he M

etho

d su

gges

ts th

at

desi

gner

s in

nova

te n

ovel

sol

utio

ns

but t

he M

etho

d do

es n

ot g

o be

yond

th

at.

• Th

e Pe

rsua

sive

Sys

tem

s D

esig

n (P

SD

Mod

el d

oes

not e

xplic

itly

reco

mm

end

that

des

igne

rs w

ork

in a

n ite

rativ

e fa

shio

n. T

he g

ener

al a

ppro

ach

sugg

este

d is

line

ar.

• Ex

plic

itly

and

at le

ngth

reco

mm

ends

ite

ratio

ns i.

e. “

Step

7: T

est a

nd it

erat

e qu

ickl

y […

] the

nex

t ste

p is

to te

st

vario

us p

ersu

asiv

e ex

perie

nces

qui

ckly

an

d re

peat

edly

. A s

erie

s of

sm

all,

rapi

d te

sts

will

teac

h m

ore

than

one

big

test

.”

8:

Doc

umen

t Con

text

ual

Influ

ence

s w

ith D

esig

n Pr

inci

ples

• Ex

plic

itly

addr

esse

s co

ntex

tual

in

fluen

ces

with

cle

ar d

omai

n sp

ecifi

c de

sign

prin

cipl

es in

the

form

of

‘lens

es’.

• Ex

plic

itly

addr

esse

s co

ntex

tual

in

fluen

ces

with

cle

ar d

omai

n sp

ecifi

c de

sign

prin

cipl

es i.

e. ‘u

se c

onte

xt’.

• Su

gges

ts th

at th

e de

sign

team

mus

t pi

npoi

nt w

hy p

eopl

e ar

e no

t per

form

ing

the

targ

et b

ehav

iour

, but

off

ers

no

advi

ce a

s to

how

to d

o th

e sa

me.

9: V

alid

ate

the

Gen

eral

isab

ility

of t

he D

esig

n •

Add

ress

es th

e ge

nera

lisab

ility

of t

he

mod

el c

laim

ing

that

it is

ge

nera

lisab

le. Y

et, t

he o

nly

proo

f of

fere

d is

the

ATM

-des

ign

case

. .

• D

esig

n ca

se a

naly

sis

is o

ffer

ed a

s pr

oof

of g

ener

alis

abili

ty. T

he a

utho

rs o

penl

y st

ate

that

the

wor

k is

con

cept

ual a

nd

theo

retic

al i.

e. b

ased

on

‘pos

tula

tes’

.

• M

entio

ns th

at th

e m

odel

is ta

ilore

d fo

r th

e ea

rly s

tage

s of

per

suas

ive

tech

nolo

gy d

esig

n bu

t doe

s no

t add

ress

ge

nera

lisab

ility

.


the ‘persuasiveness’ of applications, we need to discuss simple features or the whole systems effect. An obvious problem is that none of these suggested tactics by themselves individually appear to be persuasive. For instance, is a tactic such as reduction (making a task easier to complete) in fact a persuasive strategy or just a prerequisite? From one perspective, the models are seemingly collages of such tactics adopted from various fields of research.

The Design with Intent (DwI) model meets four of the Principles: 1, 3, 8, and 9 (but with little proof). It fails to adhere to Principles 4, (considering that the model itself was developed iteratively) 5, 6 and 7 (again, the model itself was developed iteratively). Principle 2 was partly met, on account of the fact that the model offers very clear research goals but it does explicitly mention or address theory development as such. With respect to TEL a clear advantage of this model is that it is very well grounded and corroborated, as it draws on many diverse sources of scientific knowledge by offering a plethora of references. With reference to persuasion in the domain of educational research, the DwI model seems to offer the most especially in situations, where one is designing a physical environment or an actual product e.g. for a classroom setting. It is very positive that the model is component based with an offset in specialist domains. It thus offers researchers the possibility to extend the model in developing new domains and context specific ‘lenses’.

The Persuasive Systems Design (PSD) model also meets four of the Principles, namely 1, 2, 3 and 8. Principle 9 is partly met, as the authors demonstrate very clear awareness issues with robustness and validation, describing the work as both conceptual and theoretical. This model leaves Principles 4, 5, 6 and 7 unaddressed. The PSD model is clearly inspired by formal approaches in Information Sciences (i.e. “software requirements”). If researchers are to address web-based e-education Learning Management Systems (such as Fronter, Blackboard or Moodle), the PSD model will certainly offer significant guidance. The many clear design principles offered are reasonably straight forward; however there is not much guidance as to the right mix of principles. B.J. Fogg, is by far the single most influential PD researcher and is generally credited with being the founding father of the field of PD. However, his Eight-Step Design Process model only adheres partly to Principle 1 and (as the only model) Principle 7. Fogg openly states that the model is the result of informal synthesis of “best practices”, condensed from his personal work experience. But, despite Fogg’s undeniably strong ethos, from a scientific standpoint, this is problematic. Fogg claims that the model can facilitate rapid prototyping in design teams and states that scientific research experiments can only begin after Step 8. Nonetheless, even if one were to experiment i.e. measure the success of the resulting design, the foundation would not be optimal from a research-based perspective. The Eight-Step model seems most appropriate if researchers wish to get quickly acquainted with PD-thinking and reasoning. In fact, the model is a good starting point, if one is not familiar with PD and wishes to grasp the concept of PD more thoroughly. Nonetheless, this model may not be the optimal choice for more advanced scientific projects.

In delegating persuasion to non-human entities, we are left with an entirely new set of challenges regardless of which models we choose. Many difficult questions arise especially regarding ‘target behaviours’. What are we really persuading someone about, when we seek to persuade someone into learning something?... To spend time on studying? If students are dealing with cognitive problems, what are we to persuade them about? ‘Learning’ might not automatically occur even if students change both their behaviour and have the right attitude; we cannot persuade students into being able, if they are not. Another general problem with technology mediated persuasion is that it is problematic both for researchers, practitioners and users to fully understand how persuasive designs persuade. When we persuade in speech or writing, we – hopefully – strive to present cogent arguments and may employ appeals such as


logos, ethos and pathos. We have models if we want to analyse such arguments (Weston 2000, Walton, Reed & Macagno 2008, Toulmin 2003), but how do physical objects or interactive systems persuade users and how can users evaluate the validity of the ‘claims’ presented by designs? None of the models – for good reasons – help us answer these problems.

Conclusion

The application of Wang and Hannafin's nine principles for design based TEL research served to shed light on various traits of the PD-models. The models each had their own strengths and weaknesses and perhaps it is not surprising, that a fledgling area of research fails to address the entire complexity offered in the nine principles. Moreover, these are certainly issues that new iterations of frameworks would somehow have to integrate and therefore, there would be several larger issues left to discuss. But a first step has been taken towards addressing the possibilities of creating a ‘learning’ centric Persuasive Design model. It is clear that a successful model would have to address an extremely complex mesh of social science concepts. When we seek to apply persuasion to a rhetorical situation centred on learning, we are faced with considerably large and diverse challenges. An initial challenge is that it would be very difficult to be a true specialist of learning, pedagogic, rhetoric and technology design, and thus the effort of creating a Persuasive Design (PD) model for TEL would appear to be an inherently multi-disciplinary task.

References

Aristotle 2010, 350 B.C.E-last update, Rhetoric [Homepage of classics.mit.edu], [Online]. Available: http://classics.mit.edu/Aristotle/rhetoric.html [2010, 28/5] .

Beyer, H. & Holtzblatt, K. 1998, Contextual Design -defining Customer-Centered Systems, Morgan Kaufmann Publishers, San Francisco.

Bitzer, L., F. (1968), The Rhetorical Situation, Pennsylvania State University Press, Pennsylvania, USA.

Chatterjee, S. & Price, A. 2009, "Healthy living with persuasive technologies: framework, issues, and challenges", Journal of the American Medical Informatics Association, vol. 16, no. 2, pp. 171-178.

Chi, P., Chen, J., Chu, H. & Lo, J. 2008, "Enabling Calorie-Aware Cooking in a Smart Kitchen", Persuasive Technology - Third International Conference, PERSUASIVE 2008, eds. H. Oinas-Kukkonen, P. Hasle, M. Harjumaa, K. Segerståhl & P. Øhrstrøm, Springer - Proceedings Series: Lecture Notes in Computer Science, , pp. 116.

Christensen, A.-., A. & Hasle, P.F.V. 2007, "Classical Rhetoric and a Limit to Persuasion", Second International Conference on Persuasive Technology, PERSUASIVE 2007, eds. Y. Kort, W. IJsselsteijn, C. Midden, B. Eggen & B. Fogg J., Springer, Germany, pp. 307.

Dillard, J., P. & Pfau, M., W. 2002, The Persuasion Handbook: Developments in Theory and Practice, 1 edition edn, Sage Publications, Inc;, Thousands Oaks, California.

Fabrique 2013, , inSights - a tool for persuasive design [Homepage of http://www.fabrique.nl], [Online]. Available: http://www.fabrique.nl/insights/ [2013, June/24] .

Fogg, B.J. 2009, "Creating persuasive technologies: an eight-step design process", Persuasive '09: Proceedings of the 4th International Conference on Persuasive TechnologyACM, New York, NY, USA, pp. 1.

Fogg, B.J. 2003, Persuasive Technology - using Computers to Change What We Think and Do, Morgan Kaufmann Publishers, San Francisco.


Fogg, B.J. 1999, "Persuasive technologies", Communications of the ACM, vol. 42, no. 5, pp. 26-29.

Fogg, B.J. 1998, "Persuasive Computers: Perspectives and Research Directions", CHI 98 Los Angeles USA - 18-23 April 1998, , pp. 225.

Futerra 2013, June, 2013-last update, Change Maker Cards [Homepage of http://www.futerra.co.uk/], [Online]. Available: http://www.futerra.co.uk/work#go=change-maker-pack-9766 [2013, 6/30] .

Holtzblatt, K., Wendell, B., J. & Wood, S. 2005, Rapid Contextual Design -a How-To Guide to Key Techniques For User-Centered Design, 1st edn, Morgan Kaufmann Publishers, San Francisco.

IJsselsteijn, W., de Kort, Y., Midden, C., Eggen, B. & van den Hoven, E. 2006, "First International Conference on Persuasive Technology for Human Well-Being, PERSUASIVE 2006,", Springer - Proceedings Series: Lecture Notes in Computer Science, Eindhoven, The Netherlands, pp. 1.

Kamal, N. & Fels, S. 2012, "Determining the Determinants of Health Behaviour Change through an Online Social Network" in , eds. M. Bang & E. Ragnemalm, Springer Berlin Heidelberg, , pp. 1-12.

Lockton, D. 2013, 2010-last update, Design with Intent toolkit wiki [Homepage of Dan Lockton], [Online]. Available: http://www.danlockton.com/dwi/Main_Page [2013, 05/14] .

Lockton, D., Harrison, D., Holley, T. & Stanton, N.A. 2009, "Influencing interaction: development of the design with intent method", Persuasive '09: Proceedings of the 4th International Conference on Persuasive TechnologyACM, New York, NY, USA, pp. 1.

Lockton, D., Harrison, D. & Stanton, N., A. 2008, "Design with Intent: Persuasive Technology in a Wider Context", Persuasive Technology - Third International Conference, PERSUASIVE 2008, eds. H. Oinas-Kukkonen, P. Hasle, M. Harjumaa, K. Segerståhl & P. Øhrstrøm, Springer - Proceedings Series: Lecture Notes in Computer Science, , pp. 274.

Lockton, D., Harrison, D. & Stanton, N.A. 2010, "The Design with Intent Method: A design tool for influencing user behaviour", Applied Ergonomics, vol. 41, no. 3, pp. 382-392.

Nike 2013, , Nike+ running system [Homepage of Nike], [Online]. Available: http://nikeplus.nike.com/plus/ [2013, 5/20] .

Obermair, C., Reitberger, W., Meschtscherjakov, A., Lankes, M. & Tscheligi, M. 2008, "perFrames: Persuasive Picture Frames for Proper Posture", Persuasive Technology - Third International Conference, PERSUASIVE 2008, eds. H. Oinas-Kukkonen, P. Hasle, M. Harjumaa, K. Segerståhl & P. Øhrstrøm, Springer - Proceedings Series: Lecture Notes in Computer Science, , pp. 128.

Oinas-Kukkonen, H. & Harjumaa, M. 2009, "Persuasive Systems Design: Key Issues, Process Model, and System Features", Communications of the Association for Information Systems, vol. Vol. 24, Article 28.

Parmar, V., Keyson, D. & deBont, C. 2008, "Persuasive Technology for Shaping Social Beliefs of Rural Women in India: An Approach Based on the Theory of Planned Behaviour", Persuasive Technology - Third International Conference, PERSUASIVE 2008, eds. H. Oinas-Kukkonen, P. Hasle, M. Harjumaa, K. Segerståhl & P. Øhrstrøm, Springer - Proceedings Series: Lecture Notes in Computer Science, , pp. 104.

Realityworks, I. 2010, [Homepage of http://www.realityworks.com/], [Online]. Available: http://www.realityworks.com/realcare/index.html [2010, 02/26] .

Redström, J. 2006, "Persuasive Design: Fringes and Foundations", Persuasive Technology - First International Conference on Persuasive Technology for Human Well-Being,


PERSUASIVE 2006, eds. W. IJsselsteijn, Y. de Kort, C. Midden, B. Eggen & E. van den Hoven, , pp. 112.

Torning, K. 2008, "Persuasive Technology Design - A Rhetorical Approach", Persuasive Technology - Third International Conference, PERSUASIVE 2008, eds. H. Oinas-Kukkonen, P. Hasle, M. Harjumaa, K. Segerståhl & P. Øhrstrøm, Springer - Proceedings Series: Lecture Notes in Computer Science, , pp. 83.

Torning, K. & Oinas-Kukkonen, H. 2009, "Persuasive system design: state of the art and future directions", PERSUASIVE, pp. 30.

Toulmin, S.E. 2003, The Uses of Argument, Updated ed. edn, Cambridge University Press, New York, NY.

Vatz, R.E. 1973, "The Myth of the Rhetorical Situation.", Philosophy and Rhetoric 6, , pp. 154-161.

Walton, D., Reed, C. & Macagno, F. 2008, Argumentation schemes, Cambridge University Press, New York, NY.

Wang, F. & Hannafin, M.J. 2005, "Design-based research and technology-enhanced learning environments", Educational Technology Research and Development, vol. 53, no. 4, pp. 5-23.

Weston, A. 2000, A Rulebook for Arguments, 3'rd edition edn, Hackett Publishing Company, Inc., Indianapolis, USA.

WHO Regional Committee for Europe 2008, 15–18 September 2008-last update, Behaviour change strategies and health: the role of health systems [Homepage of WHO], [Online]. Available: http://www.euro.who.int/__data/assets/pdf_file/0003/70185/RC58_edoc10.pdf [2011, 30/12/2011] .


Immersive Layers Design Exploring Culture

through a Persuasive Multimodal Interface Christian Grund Sørensen

Center for Computer Mediated Epistemology, Aalborg University, Denmark

[email protected]

The aim of this paper is to uncover possibilities for creating an interface design that

supports and enhances users in accessing content in a cultural mediation context. Context

is the EuroPlot-project (an EU-supported research project under EACEA) and in

particular the Kaj Munk Case. This case focuses on the dissemination of biography,

works, and impact history of Danish theologian, journalist, and playwright Kaj Munk.

In the course of developing relevant technology enhanced learning aids and identifying

persuasive learning designs it has become evident that challenges arise in the areas of

categorization, navigation and handling of complexity. It is also clear that challenges of

persuasion and immersion are a vital part of the communicative and reflective process.

It is thus necessary to provide designs of overview, persuasiveness, and immersion.

This paper suggests an Immersive Layers Design as a helpful tool. It is based on a

geographical layer, a temporal layer and a conceptual layer. These layers present access

keys to multimodal, digital content. Layers are dynamic and designed to facilitate

tangible functions to support intuitive cognition. The possible benefits of this approach

are discussed and several questions for further research are suggested.

This paper is mainly inspired by the insights of persuasive technology applied to learning

(Fogg, Gram Hansen et.al.), learning (Biggs & Tang, Bloom et.al, classical rhetoric

(Aristotle, Hasle et.al.), and conceptual assessment interface (Sørensen & Sørensen).

Keywords: Immersive Layers Design, persuasive technology, interface, intuitive

Introduction

In recent years digital resources have been made available for the public in a multitude of

different contexts. The apparently omniscient Internet provides most likely the most

comprehensive knowledge base gathered in history proving even voluminous encyclopaedias

pitifully unaccomplished in comparison. The Internet, however, would not possess this

richness and diversity, be it not for the content supplied by multifarious providers.

This is not least true in the area of cultural mediation. In a hypercomplex society (Qvortrup)

information is legion, however often also limited to dissemination in autopoietic subsystems

(Luhmann). Museums and cultural institutions face the challenge of reaching out to a broader

audience as the focus of these institutions is transcending from research and archival activities

to presentation and dissemination as the rationale for the economical investment (Falk,

Hooper-Greenville). The concept of public service thus calls for a contemporary and

medialogically relevant mediation of cultural content as is required in several contexts.

A challenge that was soon uncovered in the EuroPLOT-project is the complexity of the

content related to Kaj Munk. It includes theatrical plays, movie, sermons, newspaper articles,

novels, lyrics, secondary literature, historical accounts, and pictures. Obviously the formats of

this content are quite diverse which does support the necessity of a multimodal retrieval

system. Accessibility is partly accomplished with the Munk Study Edition of searchable,

annotated texts (Petersen & Øhrstrøm). A variety of other learning aids and content are also

available and an increase in multimedia learning aids and content should be expected in the

years to come.


Presenting such complex information in a digital environment does, however, give rise to

several challenges for the optimal exploitation of this multimodal content:

1. Categorization – How may content be categorized as to support cognitive processing?

2. Navigation – How may content be simple and intuitively accessible for users?

3. Reduction of complexity – How may content be structured to enhance learning?

If the content presented is of a nature that calls for existential reflexion several further

concerns may be added:

4. Persuasion – How may persuasive principles and design optimize utilization?

5. Immersion – How may content presentation and architecture support reflective learning?

Obviously, this is merely a tentative list of initial issues that will be addressed here. The aim

of the paper is to suggest an approach to an Immersive Layers Design facilitating easy and

cognitively tangible access to complex and multifaceted content in the realm of culture.

Suggestions in this paper are generic though the examples are specific, rooted in the

dissemination of the life and works of Kaj Munk. At the same time this paper reflects on the

role of such an interface in a persuasive principle design process in line with a persuasive

learning design network.

Background The Immersive Layers Design – concept (ILD) is developed in context of the EuroPLOT

project (PLOT = Persuasive Learning Objects and Technologies), which is being funded by

the Education, Audiovisual and Culture Executive Agency (EACEA) in the Life-long

Learning (LLL) programme (2010-2013).

Part of this project is focused on the provision and dissemination of cultural content, the Kaj

Munk Case, which has been explored by Aalborg University and the Kaj Munk Research

Center. Kaj Munk (1896-1944) was a leading theologian, playwright, journalist and writer.

His works commented the political and philosophical issues of the era as well as existential

and religious subjects in a modern paradigm. Being executed by German authorities in 1944

his story adds to the history of Danish WWII resistance.

The mediation of Munk’s works and his significance is in the EuroPLOT-project done

through The Munk Study Edition (Petersen & Øhrstrøm). This application is based on an

EMDROS-database (Petersen 2004) offering fully searchable annotated texts of Munk’s plays

and sermons. The Munk Study Edition is supplemented with GLO’s (Generative Learning

Objects) (Hansen, Hansen, Sørensen, Øhrstrøm 2013). A tool has been developed with a

specialized focus on assessing amongst other things cultural mediation: The Conceptual Pond

(Sørensen & Sørensen).

Concept of Immersion

An immersive, digital environment may be a transforming factor in the mediation of culture.

The concept of immersion, and the terminology, is inspired by Janet Murray’s definition in

“Hamlet on the Holodeck”. Here the concept of immersion is linked to the narrative potential

of digital, multi-media storytelling and is expressing the full absorption into a new, digital

dimension. Immersion is described as “entering the Enchanted Place” (Murray p.99) and

connotations related to arousing interest, pleasure and cognitive/emotional engagement are

intuitive.

In the concept of Immersive Layers Design the term immersion acts as an adjective defining


the overall aim of the learning process as supported by the interface. In Murray’s definition

immersion is a step to achieving two further steps of engagement, “agency” and

“transformation”. For the use in this paper no particular discrimination is made between the

various steps in the immersive process. The importance of the concept is connected largely to

the intended learning outcome rather than providing a framework for assessing the

progression of engagement. The assumption is though, that some dynamics in an immersive

process are vital.

Figure 1. Intended immersive processes

Glance Overview

Simple Complex

Manifest Conceptual

Beginner Expert

Initial interest Immersion

As displayed in figure 1. the intended immersive processes are based on an ongoing

progression in several areas of cognition leading to immersion, which is the defined gold

standard. This model may not be comprehensive but should account for the dynamics in the

creation of a cultural mediation system aiming at immersion.

From a traditional learning perspective similar general progression structures apply. Bloom’s

6-step taxonomy (Bloom) as well as the later 5-step SOLO model (Structure of Observed

Learning Outcomes) (Biggs & Tang) define intended learning outcomes and processes that

despite different terminologies are hardly contradictory to the immersion concept of ILD.

Rhetorical design

The Immersive Layers Design approach is based on the Ciceronian Aptum-model depicting

basic aspects of a communicative situation. The structural potential of a basic rhetorical

framework has intuitive advantages and mirrors fundamental structure in persuasive

communication (Hasle).

Figure 2. The Aptum model

1. Orator represents the initial communicator. In a complex technology enhanced learning

system This definition is not unproblematic. The ILD-system must facilitate access to content

from a variety of contributors, including the agile implementation of user generated content.

In the Kaj Munk Case contributors could be the Kaj Munk Vicarage Museum, the Kaj Munk

Research Center, facilitators of geocaching, teachers, contributors of digital content etc.

2. Scena defines the actual or intended receivers of the communicative content. Since the

intended audience of cultural mediation from cultural institutions is usually supposed to be

comprehensive due to public service considerations it is a substantial challenge in the Munk


Case to address adequately. Also considerations must be taken to cognitive styles and learning

strategies (Riding & Rayner).

3. Res defines the material content and matter of the communication. The life, works and

impact history of Kaj Munk is a comprehensive subject that requires delicacy in discourse and

abundance in content. The subject is related to many sub-subjects of eg. historical importance.

4. Situatio marks the stage, on which the communicative act is taking place. In ILD this is

obviously in front of a screen. The interface may, however, be used individually, at a

museum, in a classroom or on a projector screen for an audience.

5. Verba is the aspect of eloquence. What communicative “words” and communicative

content is appropriate? Obviously, the variety of media related to the Munk Case interprets

verba in a multimodal, multimedia understanding.

The implementation of the Aptum model (Aptum = the appropriate, Greek) for Immersive

Layers Design may be criticised for lacking a scope of interactivity. In the context of this

paper interactivity is defined as a property to the relevant rhetorical parts.

Kairos design

The threefold structure of the present Immersive Layers Design is inspired by the rhetorical

concept of kairos. Acknowledging three aspects of kairos:

1. The opportune moment: Content is mediated at the contextually appropriate time.

2. The opportune location: Content is presented in suitable (virtual) geographical

surroundings.

3. The opportune manner: Content is mediated through the appropriate technology (or the

omission of technology).

This understanding of kairos as an important factor in learning and immersion is to some

extent implemented in the division of a temporal, a geographical, and a conceptual layer. A

focus on kairos is an important factor in designing the learning system. Suitable material

should be presented at the opportune moment, at the opportune location and in an opportune

category of mediation.

Persuasive design

A third design principle links to the Functional Triad (B.J.Fogg). This simple, triangular

model suggests three fundamental categories of applied interactive digital media.

1 Tool – increases capability

2. Medium – provides experience

3. Social actor – creates relationship

All three functions obviously have a potential in the Kaj Munk case. The tool function is

important for the learning system to facilitate learning. The user should be equipped with a

toolbox of knowledge, skills and procedural insight.

The experience aspect is no less important. Especially at the initial stages of the immersive

process experience is a vital factor. The use of visual media, sounds, videos, quests, applies a

more concrete experience to the Kaj Munk learning environment. Experience is also

important in the immersive process of personal reflection.

The social actor element should be explored and integrated further in the Immersive Layers

Design, especially possibilities of portfolio working, peer sharing and social media

connectivity.


Immersive Layers Design

On the basis of these introductory yet fundamental concepts and definitions this paper

suggests an Immersive Layers Design centred around three different layers, each supporting

overview and exploration. In casu EuroPlot the three layers function as a joint interface for

the immersion into the time, life, works, and impact history of Kaj Munk.

Each layer has its own specific logic in categorizing and presenting content. Obviously some

content is more intuitively accessible through one specific layer than through the other layers.

Other content may be equally intuitively accessible through all layers.

In Table 1. each layer is labelled and commented with a number of properties being

significant to the particular layer. Some overlap is natural regarding the properties since layers

are designed to support user interaction and immersion and are not directly linked to the

underlying categorization of content and content mediation style.

Properties are sectioned into general properties, that depict obvious characteristics of the

particular layer, and persuasive properties, that refer to various functionalities supporting

persuasiveness. This persuasiveness entails some reference to the terminological framework

of BJ.Fogg but basically supports a more open approach to persuasive principles in learning

like the position argued in the EuroPLOT Persuasive Learning Design Framework (Hansen

WP3.1).

An important element og Table 1. is the depiction of certain learning values. Again, some

overlap may be natural. Nevertheless it is imperative that certain intended learning outcomes

or values are addressed in the design of the individual layer.

Table 2. Immersive Layers Design

Immersive layers design Layer 1 Layer 2 Layer 3

Description Geographical Temporal chronological Conceptual

General properties Visual, maps, pictures Timeline, time shudder Colour/picture based

Persuasive properties Explorability, virtual

locationbased, reducing

level of abstraction,

Explorability, simulation,

interaction,

Intuitive access,

overview, reduction,

handling of complexity

Learning value Reducing level of

abstraction, acquiring

understanding of

geographical context and

real-life events

Recognizing causal

events, exploring

timelines, reference to

world outside the Munk-

universe

Acquiring overview,

supporting reflection,

facilitating deep access to

knowledge systems,

linking content to other

layers

Geographical layer interface

The interface is designed with a focus on reflective experience as well as navigation. Since

geographical positioning or location is an intuitive point of approach for many users in many

contexts a visual presentation of this should facilitate cognitively appropriate overview.

Obviously the geographical layer may in some contexts be less relevant than other layers.

This is a context dependent question depending on the actual importance of location. In a

cultural mediation context as is the ecology of this paper location presumable is important,

and in the Kaj Munk Case this is definitely of substantial relevance.

In the Kaj Munk Case the geographical layer utilizes a historical map of Vedersø, the village

and rural area in which a great number of incidents related to the story of Munk are situated.

To support an intuitive historical immersion image is an aerial photo taken by Luftwaffe

during the last year of Munk’s life,1944. The obvious provenience of the photo supports the

war-like connotations and an emotional experience.


Aerial zoomable out into a map of Denmark focusing on other areas relevant to the life and

work of Munk as well as a map of Europe. The less detailed map of Europe is important in

facilitating understanding of the historical development of fascism being a major source of

inspiration for Munk in many of his plays and leading up to WWII.

The geographical layer interface offers support for exploration. This may contribute to

reducing the level of abstraction thus providing a more relevant learning experience.

Temporal layer interface

Maps are designed to be zoomable on a temporal axis as well.

Moving dynamically from historically contemporary mapping to

contemporary mapping of present day. Overlying time layers

(maps/aerials) support a virtual time-travel. This supports the

immediate understanding of coherence between historical periods.

To mediate this understanding of chronology and kairos and a

time-line shudder controls morphing as well as presenting

contemporarily relevant pictures and e.g. WWII video clips in the

periphery of the interface. Sounds support the visual experience.

The shudder allows interactive movements back and forth in time

supporting an intuitive comprehension of development in historical events.

Supporting interactivity maps are zoomable with access to multi-media experiences

supporting the experience of independent immersion into the learning environment. This is

done by providing abundance, so the user experiences persuasion through the element of

choice.

Figure 3. Temporal layer dynamics (reduced)

Overview: Map Aerial 1944 (Luftwaffe) Airial 1945 (RAF) Maps/photos Google Maps

Content: Fascism Plays Sermons WWII Biography Impact history Contemporary reflection

Figure 3. displays a very simplified concept of the timeline. The temporal layer may be regarded as a

dynamic version of the geographical layer adding temporal flow and temporally relevant content to the

overview. The special learning relevance lies in the causal events imaging and timeline exploration.

Conceptual layer interface

The conceptual map level facilitates browsing through subjects, ideas and works. This layer

leaves the tangible metaphors of location and time. Nevertheless, it is important to support

and emphasize a concentration on ideas and discourse as this is a vital factor in immersion as

well as an often less recognized part of understanding a cultural subject in depth.

Figure 4. displays a simplified model of the conceptual layer interface. A number of key

issues and concepts in the works and thinking of Kay Munk are displayed in a way that

privileges content relevance and conceptual neighboring . Categorization is thus done with a

focus on issues and concepts featuring in the same discourses of Munk rather than a more

formalistic categorization.

In the actual interface design coloured areas are designed to merge neighboring issues and

concepts in a pursued seamless interface. Colours are omitted in this reduced model to

enhance overview.


Figure 4. Conceptual layer overview (reduced)

A relevant example could be that of a user approaching the existential issues treated in

Munk’s play The Word (1925) (marked with a dotted circle). A number of issues such as

Christianity, atheism, faith, miracle etc. are of evident relevance and a simple mouse-over

procedure may highlight links from this interface into the archival and learning material of the

complete compilation as well as external links.

Using this interface should support the user in finding not only information about relevant

concepts and issues bit even more acquiring an impression and understanding of the

connection between the individual concepts and issues with the intention of facilitating a

deeper level of immersion.

Conclusion

Much good work is done in the area of producing content for digital mediation of culture. In

the Kaj Munk Case this includes fully searchable databases of plays and sermons. It includes

geocaches on site. It includes smaller, learning objects in the shape of Generative Learning

Objects. It also includes a vast number of articles, pictures, video footage, background

information etc. The challenges of categorization, navigation and reduction of complexity

remain all the same.

This paper aims to facilitate a new kind of overview and tangible user interaction by

suggesting the Immersive Layers Design as a framework for supporting such cultural

mediation. This is done through supporting and enhancing immersive processes with the user

through simple, persuasive principles. Introducing kairos as a framework seems to benefit the

simplicity and persuasiveness of the Immersive Layers Design.

Learning contexts linked to temporal sequence do not necessarily need a special interface for

overview (Sørensen). In case of complex learning environments featuring comprehensive

learning material and multimodal technology enhanced learning some kind of persuasive

interface appears needed in order to reduce complexity.

Immersive Layers Design is a suggestions for such an interface.

Further research

A model the Immersive Layers Design needs to be tested in relation to different learning

contexts. A refinement would also be helpful especially in the field of concept categorization.

The core of this categorization is a delicate content annotation being sensitive to not only a

phrase, a picture or a piece of multimodal content in itself – bur also to the neighboring

relevance to other issues and concepts.

A further field of study is whether this Immersive Layers Design is relevant in a cultural

mediation context only, or whether some of the insights may be stretched into other learning


environments. In any event, it is evident that the close relationship between media content and

interface is a vital issue for future exploration. Amount of cultural content is rapidly growing.

Finding the right and helpful and way through the digital labyrinth is becoming even more of

a challenge.

References

Behringer, R., Gram-Hansen, S.B., Soosay, M., Mikukecka, J., Smith. C., Winther-Nielsen, N,. and Herber, E.,

(2013) Persuasive Technology for Learning in Business Context. ASCENT 2013 Conference, Kuala Lumpur,

11.-12.April 2013.

Biggs, J & Tang, C. (2007) Teaching for Quality Learning at University (3rd edn) Buckingham: SRHE and

Open University Press

Bloom, B.S.(1956) Taxonomy of Educational Objectives, the classification of educational goals – Handbook I:

Cognitive Domain , New York

Emdros. Emdros - the database for analyzed and annotated text. http://emdros.org/ (2013)

EuroPLOT: Persuasive Learning Objects and Technologies. http://www.eplot.eu/ (2010)

Falk, J. (2004) (The Director’s Cut: Toward an Improved Understanding of Learning from Museums, Sci. Ed.,

88: S83–S96. doi: 10.1002/sce.20014

Fogg, BJ (2003) Persuasive Technology - Using Computers to Change What We Think and Do. 2003, San

Francisco: Morgan Kaufmann Publishers

Gram-Hansen, L. B. (2009) “Geocaching in a persuasive perspective”, in “Proceedings of the 4th

International Conference on Persuasive Technology”, ACM New York 2009

Gram-Hansen, S.B., Scharfe, H, Dinesen;J (2011) D.3,1 Persuasive Learning Design Framework 7/2011

Europlot deliverable

Gram Hansen, S.B., Gram Hansen, L.B., Sørensen C.Grund,, Øhrstrøm, P Deliverable 5.3 PLOTLearner v3: The

Kaj Munk Case

Hasle.P., Christensen, A.K.K (2007): Classical Rhetoric and a Limit to Persuasion. Persuasive 2007, LNCS

4744. Springer

Hasle, P. (2006) The Persuasive Expansion - Rhetoric, Information Architecture, and Conceptual Structure, in

ICCS 2006, H. Schärfe, P. Hitzler, and P. Øhrstrøm, Editors. 2006, Springer

Hooper-Greenhill, E. (2004): Measuring Learning Outcomes in Museums, Archives and Libraries: The Learning

Impact Research Project (LIRP), International Journal of Heritage Studies, 10:2, 151-174/2004

Murray, Janet (2007) Hamlet on the Holodeck: The Future of Narrative in Cyberspace, New York 1997

Riding, R. & Rayner, S (2000) Cognitive Styles and Learning Strategies., London

Petersen, Ulrik (2004). Emdros - A Text Database Engine for Analyzed or Annotated Text. In: ACL, COLING

2004 Geneva, 20th International Conference on Computational Linguistics, Volume II. Proceedings, pp.

1190-1193.

Sandborg Petersen, U. & Øhrstrøm, P. (2010) Det Virtuelle Forskerværksted for Kaj Munk Studier in

Munkiana, 44/2010

Grund Sorensen, C., (2013) Easter on Twitter - a Digital Pilgrimage on the Mobile in “Viva Vox Evangelii”,

Evangelische Verlagsanstalt. Leipzig

Grund Sorensen, C. & Grund Sørensen, M. (2013) The Conceptual Pond – a Persuasive Tool for Quantifiable

Qualitative Assessment in “Emerging Research and Trends in Interactivity and the Human-Computer

Interface”, IGI Global

Tørning K., Oinas-Kukkonen H. (2009) Persuasive System Design: State of Art and Future Directions. ACM

International Conference Proceeding Series, Vol. 350


A Unified Framework For Analysing, Designing And Evaluating Persuasive Technologies

Isaac Wiafe School of Technology

Ghana Institute of Management and Public Administration Achimota, Accra - Ghana

[email protected] Abstract: Methods for designing persuasive systems continue to be a challenge and although some frameworks have been proposed to remedy this, they are all characterised with limitations that impedes their effective usage. In light of this, we propose the Unified Framework for Analysing, Developing and Evaluating persuasive technologies (U-FADE), which integrates key concepts in the Persuasive Systems Development model, the 3-Dimensional Relationship between Attitude and Behaviour model, Persuasive Pathway Model and Persuasive Technology Onion to address this limitation by providing a stepwise approach to facilitate persuasive technology design. The proposed integrated framework is conceptual and thus there is the need for further studies to investigate and demonstrate how it can be put to use. Keywords: Systems design, behaviour change, models and frameworks, strategies

Introduction

Persuasive Technology (PT) designers are faced with issues regarding design methods. This may be due to the fact that it is considered as a relatively new domain in Information Systems (IS) development. Methods such as the Persuasive Systems Design, (PSD) Model (Oinas-Kukkonen and Harjumaa, 2009), the Behaviour Wizard (Fogg and Hreha, 2010), the and the Persuasive Tool Kit (Saidin, 2011), fail to provide all the relevant information needed for designing PT applications. This impedes their effective usage. Mostly, these frameworks do not consider the changing needs of users, and persuasive systems or applications that employ these methods become obsolete with time.

Although one can argue that the 3-Dimensional Relationship between Attitude and Behaviour (3D-RAB) model (Wiafe et al., 2011b) attempts to cater for issues regarding changes in users, it also failed to provide a holistic approach to persuasive application design. This study therefore proposes the Unified Framework for Analysing, Designing, and Evaluating persuasive systems (U-FADE). The framework seeks to address some of the existing limitations of PT design approaches by bringing together and modifying key concepts from various methods. Particularly, it provides steps and tools that can be used at each stage during the design of PT applications.

Due to space constrains, the paper starts with the general overview of U-FADE followed by discussions on the various stages of the framework. At each stage, the theoretical underpinnings for the suggested tools are discussed. This is followed by discussions on practical implications of U-FADE and a conclusion.

Analysing, Designing and Evaluating Persuasive System

The U-FADE borrows it key principles and concepts from the PSD model. The PSD model is one of the most comprehensive frameworks for developing persuasive systems. It consists of


three main phases that is based on the principles that before a successful persuasive system is designed, the designer needs to i) understand key issues behind the development, ii) analyse the persuasion context, and iii) ensure that the system demonstrates a specific set of qualities or features. Oinas-Kukkonen and Harjumaa (2009) claimed that seven postulates needs to be considered during design. Two of these relates to the designer’s assessment of the user, two on persuasive strategies, and three on the system features. Although, the PSD model presents an interesting approach to PT design

designers may find it challenging in applying it effectively. This is because the information it provides is not sufficient to guide the designer on how to perform the activities mentioned in the model. For instance, even though it argues that the selection of appropriate routes, methods and strategies is necessary for designing an effective PT, it fails to provide information on how to select routes, massages and strategies. In addition, the PSD model assumed that once a system is fully functional or operational, behaviour change would be achieved. This is because it does not consider the evaluation of behaviour and how to readjust the system to cater for changing needs of the user; although Oinas-Kukkonen and Harjumaa (2009) acknowledged the need to incorporate changes in user needs. Räisänen et al. (2010) argued that although the concepts applied in the PSD model are useful, it would be more appropriate to incorporate other models into it. In this regard, we proposed the U-FADE (see figure 1), a framework that expands the PSD model to guide designers on how to perform analysis by providing tools at each stage. U-FADE consists of three main layers. The first layer presents the five stages involved in designing PT which are: Event Analysis, the Persuasive Strategy, Systems Features, Development and Implementation and Behaviour Change evaluation. The second layer consists of the various activities and outcomes that are expected to be perform at each particular stage whereas the innermost layer suggests tools that can be used for the suggested activity. In total U-FADE suggest that four main tools are useful for the design of PT applications. Below is a discussion on the various stages of the model.

Event Analysis The initial stage in U-FADE is Event Analysis. It consists of Use Context and User Context analysis. As explained by Oinas-Kukkonen and Harjumaa (2009) Event analysis enables designers to analyse issues which are peculiar to the problem domain. It highlights the internal and external factors that affect the individual to change or maintain a behaviour. U-FADE proposes that two tools are useful at this stage: the 3D-RAB model for User Context analysis and the Persuasive Technology Onion (Wiafe et al., 2011a) for Use Context analysis.

Figure 1: A Unified Framework for Analysing, Designing and Evaluating persuasive technology (U-FADE)


User context In User Context, the 3D-RAB model facilitates analysis by categorising users into 8 states based on variations in their cognitive dissonance. It provides information on target users in terms of their Current Behaviour (CB), Attitude Towards Target Behaviour (ATTB) and Attitude Towards Changing/Maintaining Behaviour (ATCMB). Wiafe et al. (2011b) defined CB as the existing action of a person in relation to the environment and ATTB as the like or dislike for a target behaviour. ATCMB measures the agreement or disagreement of a person in relation to a change in a negative behaviour or maintenance of a positive behaviour. A user or a potential user of a PT application is therefore considered to be in one of these states at any time by assessing the parametric permutation of values (positive/negative) for each of the three factors. It argues that persuasion is achieved when users move from an undesirable to a desirable state. State 1 is considered to be the most desirable state with all factors being positive and state 8 is the worse cases state with all factors being negative. Different issues present different challenges and improper design analysis may not factor the target user’s evaluation of the issue at stake. As compared to other models, the 3D-RAB model focuses on intrapersonal factors of persuasion. These are factors which may serve as internal motivators or impediments for a user to change his or her behaviour (Aronson, 1997). Specifically, the 3D-RAB model expands the cognitive dissonance theory (Festiger, 1957) for the purpose of PT design. Use context The information obtained from User Context analysis serves as an input for the Use Context analysis. The Use Context addresses issues which are domain specific. As PT focus on methodologies and tools for developing persuasive systems in addition to studying the organisational, social and end-user impact, it should incorporate the analysis of surrounding agents and systems which have direct and/or indirect impacts on its success. In particular, it should identify how changing certain key activities or personalities within the community or the target group can facilitate a particular change in behaviour or attitude (Wiafe et al., 2011a). This is because PTs are expected to operate alongside other behavioural change activities within the society. Thus it is imperative for designers to consider these issues and factor them into design. U-FADE proposes that the Persuasive Technology Onion, PTO

proposed by Wiafe et al. (2011a) facilitates the Use context (external) analysis. Informal activities of change in relation to the dissonance state of the user (identified during the User Context) should be studied first. This is the Natural Attitude or Behaviour Change, NABC (Wiafe et al., 2011a). NABC consist activities within the immediate environment that changes the user’s behaviour or attitude naturally. This can be motivated by cultural, social, norms, beliefs, ethics, commitments, values, etc. Although these issues are generic in terms of all IS design, it is particularly important for PT

Table 1. Relationship in 3D-RAB adopted from Wiafe et al. (2011b)

Stat

e

CB

ATT

B

ATC

MB

Stab

ility

Targ

eted

st

ate

1 + + + Stable (+) 1 2 + + - Unstable (+) 1 3 + - + Unstable (-) 1 4 + - - Unstable (-) 2 or 3 5 - + + Unstable (+) 1 6 - + - Unstable (-) 2 or 5 7 - - + Unstable (-) 3 or 5 8 - - - Stable (-) 4 or 6 or 7

1

36

8

74

52

Promote

Prevent

Figure 2: The 3D-RAB model (Wiafe et al. 2011b)


design because the main objective is to change behaviour or attitude and NABC analysis facilitates the identification of issues which are within or beyond the control of the designer. Accordingly, a designer can identify possible issues that affect target users. Issues concerning the physical appearance of the environment, the existing culture, behaviour of peers etc. all needs to be considered. Due to the complex nature of behaviour change analysis, there is no single specific method for collecting all information regarding natural behaviour or attitude change. This is because each situation presents different challenges in obtaining such information. Thus the designer must use various techniques to ensure that adequate and necessary information are gathered. Methods may include informal interviews, observations, previous experiences, etc. Another form of information that the designer needs to know about the user is the Planned Attitude or Behaviour Change (PABC). This elaborates on existing methods which are being used to change the users behaviour intentionally (this includes automated and non-automated approaches) and it may or may not include the targeted change. Most organisations, associations, employers, government legislature, among others define rules and regulations for their members to ensure that they behave in a particular manner. Such planned activities may promote or impede the target behaviour and there is therefore the need to identify and plan accordingly. For instance, in smoking cessation this type of analysis may provide information on target users who may be using smoking for stress management, since some individuals smoke to manage stress (Kassel et al., 2003). As compared to NABC, PABC is dominated by procedures that are “well-defined”. It provides how formal rules have been outlined to enable people to behave in a particular manner. It is emphasised that the distinction between PABC and NABC overlap in some cases. Hence what may be considered as a PABC factor in one context may be an NABC in another context. However, the essence is to identify the external factors that affect the user at a particular targeted state thus a swap in factors would not affect the analysis.

Persuasive Strategy The persuasive strategy considers how to craft the message (message pathway) and what technology should be used as a channel for persuasion. In applying the Elaboration Likelihood Model (ELM) (Petty and Wegener, 1999) to the 3D-RAB model, Wiafe et al. (2011b) proposed the Persuasive Pathway Model (PPM) (Wiafe et al., 2012). In U-FADE, PPM serves as a tool for analysing the persuasive strategy. It facilitates the identification of possible paths, selection of appropriate paths and selection of systems features. This is inspired by the fact that elaborated routed messages are more useful in inducing attitude change. An individual will careful evaluate the merits of a message before accepting or rejecting it, whereas peripheral routes present cues and heuristics that affect one’s behaviour without necessarily changing attitude (Petty and Cacioppo, 1984).

Identifying possible persuasive paths In PPM, transitions to states that involves attitude change focuses on using elaborated messages since they consist of logical arguments. Elaborate routes are effective for long term

Figure 3: The Persuasive Pathway Model (Wiafe et al. 2012)


attitude change whereas transitions to states that involve behaviour change need to use peripherally routed messages; as they are simple and effective in changing behaviour. Figure 3 is the PPM showing types of routed messages that can be applied to facilitate a change in state. The content of the message in each route may vary depending on the intention of the message (either targeting ATCMB or ATTB). For instance although transitions from state 8 to 7 and 7 to 5 will both use elaborated routed messages, the content of message in 8 to 7 needs to focus on encouraging attitude towards change whereas that of 7 to 5 needs to focus on attitude towards behaviour. Selecting appropriate persuasive paths The order in which persuasive messages are presented is relevant (Wiafe et al., 2011b), thus the need for selecting appropriate paths. However, although elaborated routed messages are useful in attitude change, they are more effective when the targeted user is motivated enough and have the ability to process the information (Petty and Wegener, 1999, O’Keefe, 2002). This is because the receiver needs to process the information by performing “internal” logical arguments before accepting the message. Consequently, presenting messages through elaborated routes to target users in state 8 for instance will be less effective. Users at this state have less interest in the target behaviour and may not be ready to indulge in issues that need detailed cognitive processing. Messages that are counter-attitudinal to the receiver evoke unfavourable thoughts and are less successful. Hence, it is appropriate to present peripheral messages first to such individuals to encourage behaviour change before elaborated messages are presented to target attitude change. As elaboration increases, the user’s application of heuristics (peripheral messages) diminishes and they turn to scrutinise the message and attempts to make meaning of it. Thus elaborated routes become useful. However, since the 3D-RAB model advocates for various paths that can be used for persuasion, designers may also consider other possible paths during design if it is envisaged that this approach is not appropriate in a particular context. The technology U-FADE considers technology as part of the Persuasive Strategy as compared to the PSD-model which advocates for technology to be study as part of the Event. Considering the selection of technology as a strategy enables the identification of appropriate technology for PT development, instead of limiting system functionality to suit a predefined technology. By this, the designer is capable of identifying the most appropriate technology for a particular problem in terms of purpose, availability, ease of use and acceptance. In cases where one is constrained to a predefined technology, it is considered to be a constraint to strategy. U-FADE therefore suggests that in selecting hardware for persuasion, designers should i) identify hardware which are familiar to targets users, ii) select those that the application can be deployed on and iii) select those that are readily available. Where the application cannot be deployed on any of the available hardware the designer should ensure that any new hardware introduced is easily accessible and also easy to use.

Systems Features Persuasion consists of a number of activities which may be overt or covert and thus it is almost impossible to identify what actually makes people change their behaviour or attitude. Yet it is imperative for researchers to identify methods that can guide the selection of system features during persuasive design. Although the system features proposed by Oinas-Kukkonen and Harjumaa (2009) all support peripherally or elaborated routed messages depending on the context and target users, some can be considered to be more appropriate for any particular route. Table 2 is lists of possible system features categorised according to the type of message they readily promote.


Table 2. System features supporting either elaborated or peripheral route

Feature Example implementation Pr

imar

y T

ask

Supp

ort

Elaborated Tunnelling Smoking cessation Website offers information about treatment opportunities after a user has taken an interactive addiction test

Self-monitoring Users monitor their daily physical activity on how much calories they burn in a day

Simulation Cigarette packs with before and after images of lung disease which are results of smoking

Rehearsal A flying simulator Peripheral Personalisation User are able to change the graphical layout of a weight management app

Reduction In Amazon user details are store and presented in their next visit to reduce personal data input to encourage purchases

Tailoring Personal trainer website provides different information to users who are beginners and those who are not

Dia

logu

e Su

ppor

t

Elaborated Suggestion Weight management app suggests alternative meals that are healthier Social Role An avatar playing the role of a social worker who advice and provide

emotional support for smokers to quit Peripheral Rewards Users compete for the most decorated virtual pet by following their diet

and physical activity plan in a weight management app Reminder A PDA health adviser reminds users to rest their wrist whilst using a PC Similarity Slang names can be used to make teenage users associate their lifestyle

with an application Liking Cartoons can be used to encourage children to read Praise A virtual coach can provide positive feedback when users perform

physical activity

Syst

em C

redi

bilit

y

Elaborated Expertise A company’s websites provides information on their core know-how Verifiability A website for quitters provides links to support the facts it displays Authority Website quotes an authority such as a statement by a UK NHS

Peripheral Surface Credibility

UK NHS logo on a weight management website

Trustworthiness Company website provide information related to its products rather than providing marketing information

Real-World Feel Company websites provides possibilities to contact specific people and ask questions or get feedback

Authority A website with statements and quotes made by state health officials

Soci

al S

uppo

rt

Elaborated Social Facilitation

In a mobile lifestyle coaching, users can be divided into groups and they can see their group members actual performance

Social Comparison

Users of an application share and compare information relating to their quitting methods

Peripheral Social Learning A shared fitness journal in a mobile app to encourage physical activity Recognition Names of successful quitters published on a quitters website Competition Users compete for highest weight loss of in a weight management app Normative Influence

Possibility of challenging friends and relatives in a weight management app

Transition description cards U-FADE proposes the use of a tool to summarise the entire analysis preformed. This is the Transitions Description Cards (TDC) (see figure 4). This summary provides specific information on how each of the transitions in the selected pathway would be implemented, and it is recommended that a TDC is completed for every transition from one state to another within the selected path. The proposed card is expected to contain information regarding the two states involved in the transition and highlights the previous state, the current state and the next targeted state. In addition it includes information on the general description of the targeted behaviour, the targeted state, the specific user description and transitions which


favours the target behaviour among others. TDC facilitates the development and implementation stage and it enables the designer to focus on requirement without referring to other documents used for analysis during the implementation stage. Figure 4 is a TDC template for transition n1 to n2, where n0 is the previous state of a user.

Transition: n1 → n2

Target Behaviour: State the behaviour that is expected of target users. This should be specific and overt and measurable. (This information can be acquired from the 3D-RAB model analysis) Assumptions: Lists all the assumptions related to the transition. Also provide information on what happens when assumption is false. (This is obtained from the Event Analysis) Constraints: List all identified factors that may impede the progress of transitions. (This is obtained from the Event Analysis) List of possible paths: List all possible paths identified from the initial user state to the targeted state. (This is obtained from the Persuasive pathway model)

Selected path: State the selected path for persuasion Previous state: n0 Mention the state prior to the current state

Description of Previous state: Describe the characteristics of the previous state and mention any known issue that may affect the transition. (This obtained form 3D-RAB model) Selected system features: List all selected system features to be used in the transition (This is obtained from system features analysis)

Selected technology: List technologies identified for the transition (This is obtained from the persuasive strategy analysis)

Comments: State any comment that is necessary in relation to the transition

Figure 4: Transition Description Card (TDC) template

Development and Implementation The development and implementation stage of the framework comprises generic features of the classic software development approaches and thus it is not discussed. However it is recommended that particular attention should be given to ensure that the approach used favours the design processes.

Evaluating Change PTs can result in changes which may be desirable or undesirable (Berdichevsky and Neuenschwander, 1999). In some cases there may be no change at all. Thus it is necessary to evaluate that actual change that occurs during the use of a PT application. This is one of the benefits of using the U-FADE framework since existing methods for design do not provide means for accessing the intended change. In U-FADE it is argued that the 3D-RAB model is useful for this analysis. Designers need to perform a “before and after” analysis to identify how target users have changed (moved from one state to another) over a period of time. The main benefit of using the 3D-RAB model for such investigations is that it is capable of identifying specific state changes which may not be overt, yet has an impact on behaviour. For instance, after a period of using a smoking cessation PT application the persuader may observe that a user who was in state 8 have moved to state 6 (change his attitude towards target behaviour, yet continue to smoke). This can inform the persuader on what to do next although an overt change was not achieved. It is recommended that an assessment plan is made to specify the frequency of behaviour change reassessment.


Design Implications and Conclusion

From the discussions above it can be observed that U-FADE overhauls existing processes for analysing and designing PT applications. The stepwise approach proposed by U-FADE and its provision of tools at each stage of the analysis is novel and useful since designers would be capable of performing analysis with informed methods. Currently, there is no existing approach or method that is used for the selection of system features, thus resulting in the ad hoc approach used by most designers. The approach suggested by U-FADE can guide designers on the selection of system features targeting either peripheral or elaborated messages, however it is emphasised that it does not guarantee the selection of the most effective method for persuasion. Rather it provides a thinking system that guides the designer during the selection process. In addition, with the provision of methods that can be used to categories users based on the levels of cognitive dissonance state and the behaviour reassessment tool, designers can automate the behaviour assessment process as part of the persuasive application. This will make it possible for the system to automatically change the persuasive message and strategy presented to users to suite their current state: an activity that is characterised with human persuaders. In conclusion, U-FADE simplifies and suggests tools for analysis and design of PT systems, however there is the need for further studies to assess the practicality and applicability of such a framework. This will provide practical and empirical evidence of its strengths and weakness.

References

ARONSON, E. 1997. Back to the Future: Retrospective Review of Leon Festinger's" A Theory of Cognitive Dissonance". The American Journal of Psychology, 110, 127-137.

BERDICHEVSKY, D. & NEUENSCHWANDER, E. 1999. Toward an Ethicsof PersuasiveTechnology. COMMUNICATIONS OF THE ACM.

FESTIGER, L. 1957. A theory of cognitive dissonance. Row, Peterson, New York. FOGG, B. & HREHA, J. Behavior wizard: a method for matching target behaviors with solutions. 5th

Internatioanl Conference on Persuasive Technology, 2010. 117-131. KASSEL, J., STROUD, L. & PARONIS, C. 2003. Smoking, stress, and negative affect: Correlation, causation,

and context across stages of smoking. Psychological Bulletin, 129, 270-304. O’KEEFE, D. J. 2002. The persuasive effects of variation in standpoint articulation. Advances in pragma-

dialectics, 65-82. OINAS-KUKKONEN, H. & HARJUMAA, M. 2009. Persuasive Systems Design: Key Issues, Process Model,

and System Features. Communications of the Association for Information Systems PETTY, R. E. & CACIOPPO, J. T. 1984. Source factors and the elaboration likelihood model of persuasion.

Advances in Consumer Research, 11, 668-672. PETTY, R. E. & WEGENER, D. T. 1999. The elaboration likelihood model: Current status and controversies.

Dual-process theories in social psychology, 41-72. RÄISÄNEN, T., LEHTO, T. & OINAS-KUKKONEN, H. Practical findings from applying the PSD Model for

evaluating software design specifications. 5th Internatioanl Conference on Persuasive Technology, 2010. 185-192.

SAIDIN, A. Z. Persuasion Knowledge Toolkit: Requirements Gathering with Designer. The 25th BCS Conference on Human Computer Interaction, 2011 Newcastle.

WIAFE, I., ALHAMMAD, M. M., NAKATA, K. & GULLIVER, S. R. Analyzing the Persuasion Context of the Persuasive Systems Design Model with the 3D-RAB Model. 7th International Conference on Persuasive Technology, 2012 Linkoping Sweden. LNCS.

WIAFE, I., NAKATA, K. & GULLIVER, S. R. Designing Persuasive Third Party Applications for Social Networking Services Based on the 3D-RAB Model. In: PARK, J. J., YANG, L. T. & LEE, C., eds. Future Information Technology, 2011a. Springer Berlin Heidelberg, 54-61.

WIAFE, I., NAKATA, K., MORAN, S. & GULLIVER, S. Considering user attitude and behaviour in persuasive systems design. European Conference on Information Systems, 2011b Helsinki, Finland.


Assessment and Learning Success

For any kind of learning, it is important to assess the success of the learning, to ensure that the

learning outcomes have been achieved and that the learner has actually learned the material that

was taught in the teaching process. For measuring this success, assessments are required, which can

be either formal or informal. This is the same in both human-taught learning as well as in machine-

controlled automatic teaching systems.

In the persuasive learning approach, also the assessment is to be based on persuasive design

principles. In their paper, Gottschalk and Winther-Nielsen investigate the use of such persuasive

concepts for monitoring students’ learning success.

Another interesting aspect in persuasive learning is to measure the actual degree of motivation and

persuasion, which Eric Brangier and Michel Desmarais have investigated in their paper.


Persuasive skill development: On computational surveillance strategies for modeling learning statistics with PLOTLearner

Judith Gottschalk, Aalborg University, Nyhavnsgade 14, 9000 Aalborg

[email protected]

Nicolai Winther-Nielsen Fjellhaug International University College Denmark, Frederiksborggade 1.B.1,

DK-1360 Copenhagen K

[email protected]

The hypothesis this paper is based on is that computational surveillance within computer-assisted

language learning [CALL] systems helps teachers to give their students detailed feedback and

enables the computational system to give response to students. We discuss this hypothesis and

give as example research work in progress on PeLLITS, short for a Persuasive Intelligent Tutoring

System, which is an extension of PLOTLearner. As a result of the data mining it is possible to

derive and model learning statistics by using the formal kernel of Item Response Theory [IRT]

which has been proposed by Rasch in 1960 and is now used in a paper by Metsämuuronen from

2013. This way we integrate a feedback-loop in our CALL system to achieve a behavior change

in the student. Surveillance and tailoring are essential to persuasive learning with our CALL

system as this customization supports the application of modern learning theories in which

communication and feedback is crucial, following the theory of leaning design formulated by

Laurillard (2012). With the implementation of IRT in our CALL system, it is possible to predict

the students' learning progress and to give the teacher the opportunity to gently guide student at

the right point and with the best feedback. We also discuss the ethical aspects. The computational

implementation of this additional module to the online application of PLOTLearner is supervised

by Claus Tøndering, developer of PLOTLearner, who is the executive programmer and trains the

lead author of this paper in implementing the model developed in this paper.

Keywords: Computer-assisted language learning, surveillance, tailoring, feedback, learning

statistics, IRT, learning progress predication, persuasion

1 Introduction

Measuring learning statistics through computational surveillance and derived via data mining

algorithms can support persuasive teaching in computer-assisted language learning [CALL] systems.

This is the hypothesis this paper deals with. We discuss how monitoring and predicting learning

progress is being used in research work in progress on an intelligent tutoring system called Persuasive

Language Learning Intelligent Tutoring System [PeLLITS] which is an extension of the database-

driven CALL environment PLOTLearner (cf. http://bh.3bmoodle.dk/).

Currently PLOTLearner is being repurposed as an online application, Bible Online Learner

(http://pltest.3bmoodle.dk/), developed by programmer Claus Tøndering. In the new project we use

logged data on learner performance from this repurposed version of PLOTLearner as input for

modeling learner statistics to be used by students, peers and teachers. The ultimate goal of the new

project is, however, to emulate the presence of an artificial tutor as a learning supervisor in a virtual

and interactive world. The development of PeLLITS is still in the beginning; its architecture has been

developed in Gottschalk (2012), where it is also described in length.

The current status of PeLLITS is: It has a web interface for facilitators and learners using

PLOTLearner which we call Learning Journey Online. Various data mining algorithms and the

formal IRT-based framework used by Metsämuuronen (2013) will enable the facilitator to in


gathering information on the learning progress of a student. This is statistically evaluated to give an

overview of the students learning progress. The system can show the learning progress in detail, using

graphs and tables. It can give advice on where the student needs further support and where he or she

should focus on in grammatical drills, and it can predict the students learning progress as shown in

the screen-shot of the Learning Journey Online in figure 1. The surveillance module collects data

when learners choose to submit their learning statistics to the system. Based on the learning statistics,

the facilitator can give a direct and informed feedback to the learners, and learners can direct their

own learning project through the feedback from the system. This way they are able to tailor their own

learning course via surveillance. The generated feedback loop will effect a change of behavior

towards improved learning skills, which is one of the persuasive functions identified by Fogg (2003).

Figure 1: Screen-shot from Learning Journey Online

The system is inspired by experimentation in a course taught for the EuroPLOT project by Nicolai

Winther-Nielsen at Fjellhaug International University College Denmark in Copenhagen. During this

course he developed and tested the 12 introductory sessions which are delivered as open learning for

EuroPLOT (http://bh.3bmoodle.dk/course/view.php?id=2 login as guest). In this classroom students

used PLOTLearner and gave feedback on their learner experience in oral discussions during class as

well as in e-mails, and the students forwarded their learning statistics for two months during this

course. Both EuroPLOT’s PC version of PLOTLearner and a new online application, Bible Online

Learner, can manage upload to a server, and our project uses these uploaded data as input for the

Learning Journey Online. Without this solution the teacher has to analyze the learning statistics by

hand when he or she wants to provide feedback for the learner. This is both a time consuming task

and it depends to a high degree on the teacher’s previous experience of how students usually develop

their language skills when making specific errors. The course design of the lecture using

PLOTLearner was inspired by the idea that the learners are self-directed.

The research questions the current paper deals with are: 1) How does a persuasive approach to

learning statistics look like? 2) What are the formal requirements on a statistics module which is

designed to support tailored learning in a CALL system via surveillance?

Choosing a formal approach to the prediction of learning progress which employs surveillance

naturally raises ethical concerns because it is possible to predict learning progress in learners. The

paper discusses the ethical dimensions of surveillance in persuasive learning.

The paper is organized as follows: Section 2 discusses how surveillance can be persuasive and argues


that a feedback-loop which is created via surveillance and interaction with the CALL and the teacher

will have a persuasive effect on the learner. Also ethical aspects are discussed in this section. In

section 3 the formal framework for a statistics module based on Metsämuuronen is laid out.

2 How surveillance can be persuasive

The principles of persuasion that inspires our surveillance module was set out by Fogg (2003) in his

introduction to Persuasive Technology. A decade of research into the role of the computer as a

persuader suggested its potential for “adaptive education and training products that tailor motivational

approaches to match each individual learner” (Fogg 2003:246). The PLOTLearner project has

focused on how exactly technology can persuade learners and change the way they act and learn (cf.

Winther-Nielsen ms b).

The development of the surveillance module aims on the one hand to be accessed by teachers and

learners by a simple web interface which will enable them to keep track of the students' learning

progress and on the other hand to assist the student in self-monitoring his or her own learning progress.

We define surveillance according to the theory of Persuasive Technology: “One party monitors the

other party to modify the behavior in a specific way” (Fogg 2003: 46), and this is a common technique

to persuade people to change their behavior. The approach chosen is uncovered monitoring; it

persuades the learners in their learning by given instant and corrective feedback relative to the practice.

As it has been explained in Laurillard (2012) and been exemplified in Gottschalk (2012), feedback is

what most users of software for computer-assisted language learning desire. Surveillance is a way to

provide this kind of feedback via detailed learning statistics, because it supports both the teachers and

the CALL system in giving detailed feedback on the students' progress.

Figure 2: Persuasive architecture (Winther-Nielsen ms a: 7)

The persuasive architecture which Winther-Nielsen (ms a: 7) has developed for PLOTLearner is

based on Fogg (2003), proposing only a slight modification of the architecture of a persuasive tool.

Winther-Nielsen differentiates three levels in the development of ability and motivation through

computer-assisted language learning. The first and simplest method to enhance ability is reduction

which is well-known from quizzes satisfying the learner’s basic need for reviewing knowledge and

memorization. A more sophisticated approach is tunneling. It proceeds in predefined learning

progression. Such a predetermined sequence enforces automated results in mastery however it also

limits freedom. The most persuasive activation of ability is achieved however via tailoring. In such a


persuasive technology the training is adjusted to the learners’ knowledge level, age, learning style,

progression, goals and other highly individual parameters which are related to vocational needs.

It is tailoring which is supported by the PeLLITS technology. Detailed feedback based on learning

statistics will make it possible to adjust the whole learning process according to the statistics results

and the specific needs of the student. By choice of the student, the teacher receives a detailed learning

profile of this learner’s previous practice. Based on this profile the teacher can adjust his or her

teaching according to the individual needs and goals of the learners. The teacher can intervene and

provide the student with feedback and advice. Since the student gets feedback by the facilitator based

on the learning statistics the student can as a self-directed learner focus on the teacher’s advice and

adjust her or his learning according to it.

The other branch of Winther-Nielsen's architecture is motivation with similar functions and set up in

a parallel track to focus on increasingly persuasive feedback. In the traditional curriculum, the first

and often only motivational factor is conditioning which is caused by fear of failing the exam, but

rewarded by the hope of acquiring a certificate. This is at the core of learning technology with simple

exercise functions and instant conditioned feedback. At the next more sophisticated level of

motivation surveillance will offer learners the choice to voluntarily share their outcomes with teachers,

fellow learners, or in an open community. Using logged data on activity, an offer to receive more

instructional help could motivate learners to improve on their learning processes. Furthermore, they

could be given the opportunity to attain a certain status as a good gamer in a competition with other

learners or to become the best helper in collaboration among peers. However, Winther-Nielsen (ms

a) assumes that the ultimate persuasive system will be based on self-monitoring as in this case self-

directed learners experience the freedom of being able to actively explore the learning environment

and plot their own personal course through the learning content. In this way they are persuaded to

plan their individual learning journey based on a visualization of progress and they can respond to

the right kind of corrective feedback. Hence, the ultimate language learning system envisioned by

Winther-Nielsen uses artificial intelligence and natural language processing to record the individual’s

processes and outcomes and measure performance on language learning tasks. This system

automatically adapts to individual learner differences and learning processes through tailoring and it

primes the student for learning tasks thanks to surveillance and self-monitoring.

It is predominately the addition of a further feedback element to the learning circle which makes the

system persuasive. In the current version of PeLLITS this is surveillance, but in the long run PeLLITS

will be enhanced and developed into a self-monitoring system using artificial intelligence as

envisioned in Gottschalk (2012).

While Fogg's approach to persuasion has inspired work on PLOTLearner and PeLLITS, the

pedagogical approach for both is mainly based on the robust theory of design for learning which has

been introduced in Laurillard (2012). It defines what makes the system persuasive. Laurillard's

framework models the interaction between learners, facilitators and the learning environment.

Laurillard's model of conversational theory is used to visualize the processes involved in learning

from an external environment of a teacher (cf. Winther-Nielsen ms b). Laurillard's approach from is

reproduced in figure 3, adding the surveillance part to her system.


Figure 3: Surveillance added to the model of Laurillard (2012: 60)

In Laurillard's (2012: 60) approach learners use their personal goals and the current organization of

learning spaces to select a desired practice which will generate learning actions on the external

environment. As Laurillard explains, learners use actions which are modeled by the teacher or even

results of own actions to modulate and build practice and capability. If a teacher is present in the

learning cycle, there is also the possibility to learn through direct communication. Explanations and

comments from the teacher or the learning environment enable the learner to develop a conceptual

organization of the learning environment. This is shown by the elements (CT) and (FT).

It is specifically this part of the learning cycle described by Laurillard which is supported by the

surveillance module. The reason is that surveillance activates a feedback circle which results in a

behavior change in the student as he or she can adjust the learning progress according to this particular

aim. This aspect of the system makes the statistics agent a tool for persuasive learning.

However, the use of surveillance in CALL raises serious concerns of ethic nature. The role of

surveillance in persuasion has been discussed at length in Fogg (2003) and in Jespersen et al. (2007).

Both note that surveillance makes it possible to change behavior, however this only works if the

students know that they are monitored. This surveillance element might raise memories of the

Panopticum introduced by Bentham and Bozovic (1995) where prisoners are permanently monitored.

The Panopticum is a prison that in its center has a tower from which the guards can permanently

monitor the prisoners but do not necessarily do so. This results in a deprivation of privacy and forces

the prisoners to behave appropriately.

While the system was intended as an enhancement of the conditions of detention because no physical

punishment is used, Foucault (1975) later criticized this approach severely and this critique is also

followed by Jespersen at al (2007). The essential problem is that the Bethamian system simply

redefines power in an inhuman way as the prisoners in the Panopticum lack any privacy due to the

constant surveillance (cf. Jespersen at al. 2007). However, one can argue from a utilitarian point-of-

view that the Panopticum has a good economic balance and does not physically hurt prisoners, and

there is a better alternative and ethically more valid. The Panopticum brings wealth to a big number

of people while its negative effects are moderate. This would also be acceptable to Foucault (1975)

who assumes that the microwords of power cannot be changed, but one needs to deal with them.

A similar argumentation is also true for the statistics module, although with the difference in stance

towards the ethical status of surveillance mentioned by Fogg (2003) and Jespersen et al. (2007): As


long as an intended outcome is friendly and supportive for the user of the software, in that it positively

enhances objectively measurable processes relevant to the user of the learning software, it can be

regarded as ethical. This implies that if the surveillance technology is uncovered and it supports

learners in acquiring desirable skills in a target language the goal can be regarded as friendly and

therefore ethical. In the end the students immolate their privacy, but receive feedback which enhances

their learning. From the utilitarian position, this is positive and regarded as ethical.

3 Formal framework

To develop a formal statistics framework for PeLLITS Judith Gottschalk is now using Item Response

Theory [IRT] following the earlier helpful work by Metsämuuronen (2013). The main purpose of IRT

is to test people and to predict the probability of a testee's response by establishing the position of an

individual testee along the line of some latent dimension. Since IRT is used in an educational

environment, the latent trait is often called ability which is set into relation to the difficulty of an item

(Partchev 2004: 5).

A first step towards applying a formal statistics approach with IRT for PLOTLearner is the

identification of items which can be tested with PeLLITS. These items include the morphosyntax of

the learning object dealt with in PeLLITS.

For each student an SQL database receives and stores the following: data on the student identity, the

test question answered by the student (this is what is called the item in what follows), whether the

student has given the correct answer, and how much time it took the student to answer the question.

Whether an answer is correct is determined by the truth values true or false. If a student does not

respond to a question, the answer is interpreted as false in this statistical framework. This is the

information the statistics module uses to determine statistical information on both the student and the

item.

For the statistical module a dichotomous Rasch-Model is used. This is an IRT model which predicts

the probability of a learner giving a certain response to some item. Within this model it is assumed

that learners have a different level of ability and items have different levels of difficulty (Partchev

2004: 9).

The simplest IRT model for a dichotomous Rasch-Model does only possess one item parameter. It is

the item response function, which gives the probability of a correct response given the single item

parameter bi and the individual ability level θj. The item-response function which contains one

parameter is given in (1) below. It is also known as one-parameter logistic function.

(1) The formula for the item-response function (Parchev 2004: 10)

The crucial part of the formula is the expression exp(θj − bi) which predicts the probability of a correct

response from the interaction between the ability θj of an individual and the item parameter bi. The

latter parameter is called the location parameter, or the difficulty parameter (cf. Patchev 2004: 11).

The question arising from this is how the ability and the difficulty of an item are determined in this

model: These are parameters, which are estimated based on different likelihood methods. The

likelihood method, for the determination of the ability parameter, also known as person parameter is

the marginal maximum likelihood [MML] procedure. It is used to estimate the distribution of means

and standard derivations of the latent ability of each students; this is represented by the θ-parameter

in each updated version of the learning statistics right after each test is executed (cf. Metsämuuronen

2013: 17). For the estimation of the difficulty Metsämuuronen used the conditional maximum

likelihood method [CLM] because this reduces the imprecision which would otherwise result from


the person parameter which is also used for the estimation of the item parameter. This is the β-

parameter.

In view of the specific test design, it is necessary to equate the tests as pointed by Metsämuuronen.

In our context the problem of different levels of proficiency for expert and novice level students does

not affect the teacher surveillance tool because all students use PLOTLearner from the very beginning

of learning Biblical Hebrew. If, however, the tool is used by intermediate or advanced learners and

intended to store a broad number of learning profiles, the factor of comparing expert- and novice level

students will play an important role.

For the current statistics database design is even more important that the test length can differ, given

the different topics covered in the course of the week. It is important that the test scores are

comparable and therefore we introduce vertical equation. By doing this, test scores from different

tests of different lengths and administered in different time frames can still be compared statistically.

Vertical equating was used for IRT modeling in the same way as Metsämuuronen:

[… ] the scores were fitted to the same scale on the basis of characteristics of IRT models,

which assume that a learner’s latent level of ability (θ) and the difficulty level of an item

(β) are identical, when certain preconditions are met (see Wright, 1968). The latent ability

of each learner can be determined in the same scale for every test as long as linked items

connect the test versions. Because of the small number of students in the experimental

group (N = 13), the only recommendable model for estimating latent ability was the one-

parameter model (that is, the Rasch model). The estimation was carried out using the

OPLM program [...] (Metsämuuronen 2013: 17).

Accordingly, Judith Gottschalk uses the different equations, the CLM and MML introduced above

for a Rasch-Model. It is used in the same way as Metsämuuronen does and is executed in the

following order: In a first step the structure of the linked items is defined. The values of the difficulty

of the linked items are exactly the same in each version, the difficulty levels of all other items can be

calibrated according to the same scale as the one used for linked items. In a second step the CLM is

used in order to estimate the difficulty level of each item which is represented by the β-parameter.

The third step uses the MML to calculate the θ-parameter. In the final step the θ-parameter of the

scores is estimated for each version. To do this, the means and the derivations of the distributions of

the β and the θ are used. This estimation results in a unique latent value which is measured on a

common scale for each observation of the scores in all versions. Metsämuuronen points out that the

success of equating depends on three different factors which are explained.

The proficiency level of the students should be reflected by the linked items. The items should

represent a sufficient range of ability and they should be neither too easy nor too difficult. In the

intervention, the linked items for the next test are selected based on previous tests. Metsämuuronen

points out that this enables the system to discriminate items that are considered not too difficult or

too easy. The items should cover the different content areas. This selection generates a short test

embedded within the main test. It is necessary that in this short test the linked items are selected in a

way which represents different content areas as widely as possible. In the equation the stable

parameters depend on the sample. All students were tested at the beginning of the intervention to

obtain a population which is as large as possible. In this way stable item parameters are acquired:

However, from the viewpoint of the population, the parameters for items measured only

in the EG are unstable. Also, due to the small population in the intervention, the values

for item “difficulty parameter” depended considerably on those students who participated

in the test. Thus it was important to get all the possible test papers from the test-takers –

even day after the test. Although the item parameters are somewhat vague, the results are

much more accurate than if only classical metrics (the proportion of correct answers) were

used in comparison. (Metsämuuronen 2013: 17)

It will be a task for future research to develop an approach to IRT which also includes the response


time in the analysis of learning statistics. The reason why it has not been used here is that the so-

called coefficient of variation which is usually used to display learning progress based on response

time is quite controversial. This is a further argument in favour of using the IRT approach of

Metsämuuronen (2013) to formally model a statistics approach for measuring learning progress in

PeLLITS and to add surveillance to the software.

4 Conclusion

Surveillance combined with tailoring is at the core of the architecture of the new persuasive tutoring

system PeLLITS which we have outlined here. The system generates a feedback-loop which can

either be given by a teacher or a CALL system and the response motivates learners to change their

behavior as an effect of the persuasive technology. Adding a formal approach based on IRT will even

make it possible to develop a framework which is so informative that it can predict learning progress,

enhancing the motivation for behavior change even more.

Acknowledgements

This research was funded by the Education, Audiovisual and Culture Executive Agency (EACEA) of

the European Commission through the Livelong Learning Program with grant #511633. We would

like to thank Claus Tøndering for his supervision on the implementation of the statistics module

described in this paper. We would also like to thank Christian Højgaard for discussions on the design

of the statistics module.

References

Bentham, J., & Bozovic, M. (1995). The Panopticon writings, London, New York: Verso. Fogg, B.J. (2003). Persuasive Technology: Using Computers to Change What We Think and Do. San Francisco, CA: Morgan Kaufmann. Foucault, M. (1975). Surveiller et punir: naissance de la prison, Paris: Gallimard. Gottschalk, J. (2012). The Persuasive Tutor: a BDI Teaching Agent with Role and Reference Grammar Language

Interface. ITB Journal, 23, 31-51. Jespersen, J.L., Albrechtslund, A., Øhrstrøm, P., & Hasle, P. (2007). Surveillance, Persuasion, Panopticum. In de Kort

et al. (Eds.), PERSUASION 2007. Berlin, Heidelberg: Springer. Laurillard. D. (2012). Teaching as a Design Science: Building Pedagogical Patterns for

Learning and Technology. London: Routledge. Metsämuuronen, J. (2013). Effect of repeated testing to the development of Biblical Hebrew language proficiency.

Journal of Educational and Psychological Development, 3(1).

http://www.ccsenet.org/journal/index.php/jedp/article/view/19582. Parchev, I. (2004). A visual guide to item response theory. Retrieved June 1, 2013, from http://www.metheval.uni-

jena.de/irt/VisualIRT.pdf. Rasch, G. (1960). Probabilistic Models for some Intelligence and Attainment Tests. Copenhagen: Danish Institute for

Educational Research. Winther-Nielsen, N. (ms a). The Corpus as Tutor: Data-driven Persuasive Language Learning. Submitted to Digital

Humanities Quarterly.Winther-Nielsen, N. (ms b). PLOTLearner as Persuasive Technology: Tool, Simulation

and Virtual World for Language Learning. Submitted to IWEPLET 2013 (http://www.iweplet2013.eu/)


A set of criteria to assess motivation and persuasion of e-learning applications

Eric BrangierUniversité de Lorraine – Metz, PErSEUs, Psychologie Ergonomique et Sociale pour

l'Expérience Utilisateurs - BP 30309 Île du Saulcy - 57006 Metz (France)[email protected]

Michel C. DesmaraisPolytechnique Montréal, Département de génie informatique et génie logiciel,

C.P. 6079, succ. Centre-Ville, Montréal, Québec H3C 3A7 (Canada)[email protected]

Motivational factors have long been a topic of study in e-learning applications and there is a strong overlap between these factors and the persuasive factors. This study intends to investigate this overlap. Which criteria are common to both, and which criteria are specific to each? The objective is ultimately to design an interactive persuasion criteria grid that is specific to e-learning along the lines of an existing grid, while taking into account the body of research on motivational factors in e-learning. To investigate these questions, we compare the persuasion grid of Nemery, Brangier & Kopp (2011) with a number of the motivational criteria of Vincente and Pain (2002), and apply both approaches to evaluate the quality of an e-learning application to induce motivation to learn.

Keywords: Persuasive criteria, E-learning, Motivational criteria, Heursitic inspection.

Introduction

Few would question the power of some games to immerse their users in a state of regular and intense usage for long periods of time. The same observation can be said of some social sites. In that respect, these applications can be considered as reaching a high score on the persuasive scale. Incidentally, they are the same persuasive ingredients that designers of e-learning applications would like to assimilate to bring learners to use these applications more intensively and effectively.

However, motivational factors have long been a topic of study in e-learning applications (see Malone, 1981, for eg.) and there is a strong overlap between these factors and the persuasive factors. This study intends to investigate this overlap. Which criteria are common to both, and which criteria are specific to each? The objective is ultimately to design an interactive persuasion criteria grid that is specific to e-learning along the lines of an existing grid by Nemery et al. (2011).

To investigate these questions, we compare the persuasion grid of Nemery et al. with a number of the motivational criteria from deVincente (2002), and apply both approaches to evaluate the quality of an e-learning application to induce motivation to learn.

The persuasion grid was shown to help experts and, in particular, novices to identify weaknesses and potential improvements to improve the persuasive features of applications. It is composed of eight criteria and 23 sub-criteria. These eight criteria are credibility, privacy, personnalization, attractivity, sollicitation, priming, engagement, ascendency.


In the field of elearning, motivational scales were developed to assess the quality of an elearning application to engage the learner. The motivational criteria set of deVincente, (2002) is one such scale that is geared towards the assessment of motivation. It contains nine criteria: control, challenge, independence fantasy, confidence sensory interest, cognitive interest, effort, and satisfaction.

Both the persuasion grid and motivational criteria will be described and compared on a theoretical basis. Multiple dimensions will be investigated: intrinsic motivation vs. persuasive technology, sense of freedom vs. freely accepted compliance, and learning efficacy vs. learning enjoyment. This comparison will set the stage to an experiment on the analysis of an e-learning application on math.

The experiment consists in analyzing a drill and practice application in college mathematics. In the tradition of ergonomic inspection, the application will be analyzed by two experts with the persuasion criteria grid and with the aim to identify strong and weak elements and to suggest improvements. The same process will be conducted with motivational criteria taken (deVincente, 2002).

The elements identified by each approach will be compared over their similarity and differences, on a qualitative as well as a quantitative perspective.

Heuristic inspection

The Nemery et al. (2011) grid relies on the general technique of inspection. To inspect the usability of a product, whether a user interface or any artefact designed to be used by some user, is to make a judgement about its ability to be effective, efficient, error-tolerant, easy to learn and satisfying. This judgement is made by experts in ergonomics or HCI. Inspections are often the method of choice to quickly target usability issues and find the proper corrections to bring to the design of an application.

The persuasive criteria gridIn the field of education, user interfaces are becoming more and more engaging in order to encourage learners towards reaching learning objectives. Ergonomic inspection has long been considered to be a part of the evaluation of design processes, persuasive technology has not yet been taken into account. Faced with the lack of validated tool in this area, a set of criteria was elaborated. Following the review of 164 papers in the field of captology, eight criteria were considered to encompass the persuasiveness factors: credibility, privacy, personalization, attractiveness, solicitation, initiation, commitment and ascendency. These criteria are grouped under static and dynamic categories (see Table 1):

Static criteria are prerequisite elements to establish a fertile context within which a dynamic process of persuasion can be launched. These prerequisites promote the acceptance of a persuading process.

Dynamic criteria are involved in a process designed to engage the user in a series of planned and ordered persuasive steps in which the temporal factor is critical. At each step of the behavioral changes, elements of the interface bring the user to commit to greater levels of engagement.


Table 1. General architecture of the eight persuasive interactions criteriaCriteria Definitions

Credibility is the ability of the interface to inspire confidence and to make the user confident in the veracity of its information. Credibility is based on reputation and notoriety.

Privacy refers to the protection of personal data and the preservation of personal integrity and security of the interaction. It also refers to protection against loss, destruction or inadvertent disclosure of this data. of rights and ensuring the confidentiality of information.

Personalization refers to the concept of customization of the interface to the needs of the user. The customization can be a greeting, a promotion, or any means to achieve a more personal interaction with the user. It may also rely on group membership.

Attractiveness is the use of aesthetics (graphic, art, design) to capture the attention of the user, to support the interaction and create a positive emotion. The animation, colors, menus, drawings, video films are designed to catch and maintain the interest of the user.

Solicitation is the first of the four dynamic criteria. It refers to the initial stage which aims to swiftly attract and challenge the user to initiate the relationship. The interface attempts by words, graphics or any form of dialogue, to suggest a behavior and induce action through minimal influence.

Initiation refers to elements of the media that allow the first user-initiated actions. The user’s attention is captured and, on his own initiative, encouraged to realize the first engaging action. The user is caught in a gradual engagement process.

Commitment means that system further involves the user in a process. Several queries and incentives regularly and gradually engage the user. The electronic media will induce more intensive and regular behavior.

Ascendency is an expression of the completion of the engaging scenario. The user has unequivoquely accepted the logic and goals of the electronic media. The interaction is characterized by induced pleasure and possibly by the relief of internal discomfort. Ascendency is closely related to the concept of immersion in the video game field and it implies a high level of repetition and regularity of interaction, and sometimes emotional involvement in the story that result in dependence and game character identification. Users develop emotional attachment and cannot envision themselves without these product, or would feel a substantive negative effect in case of loss

The motivational criteria gridAs mentioned in the introduction, motivational factors in e-learning systems have been the topic of decades of research and we can expect an overlap with the persuasive criteria. We chose the criteria of deVincente (2002) as good representatives of the results of this research. Table 2 summarizes the motivational criteria. They are largely inspired from previous work on what makes an e-learning application engaging (see Tràn, 2008). We return to their theoretical groundings and to the distinctions and similarities that we can envision between them and the persuasive criteria in the discussion.


Table 2. General architecture of the motivational criteriaCriteria Definitions

Control Degree of control that the student likes having over the learning situation (i.e. does he like to select which exercises to do, in which order, etc. rather than let the instructor take these decisions?).

Challenge Degree that the student enjoys having challenging situations during the instruction (i.e. does he like to try difficult exercises that represent a challenge for him?).

Independence Degree that the student prefers to work independently, without asking others for help (i.e. does he prefer to work on his own, even if he finds some difficulties, and try to solve them by himself rather than asking for collaboration or help from others?).

Fantasy Degree that the student appreciates environments that evoke mental images of physical or social situations not actually present (i.e. does he like the learning materials being embedded in an imaginary context?).

Confidence Refers to the student’s belief in being able to perform the task at hand correctly.

Sensory interest Amount of curiosity aroused through the interface presentation (i.e. appeal of graphics, sounds, etc.).

Cognitive interest

Refers to curiosity aroused through the cognitive or epistemic char- est acteristics of the task (i.e. regardless of the presentation issues, does the student find the task at hand cognitively

Effort Degree that the student is exerting himself in order to perform the learning

Satisfaction Overall feeling of goal accomplishment (i.e. does the student think that the instruction is satisfying and that it is getting him closer to his goals?).

A case study for the two criteria grids

In an effort to better assess the overlap and the uniqueness of the persuasion and motivational criteria grids, we analyse an e-learning application with each of grid. The application is designed for the purpose of drill and practice on the topic of college mathematics. We will refer to it as the Exerciser. It aims to help newly enrolled engineers to assess their level of mastery of college math with respect to the level expected in their first year. If their mastery is lacking on any of topic, or if they want to enhance their skills, the Exerciser contains over 1000 problems and the equivalent of approximately 150 pages of notes that cover the theory. The notes were contextually accessed within the exercise section, and vice-versa.

Figure 1 contains some of the screen dumps of the Excerciser that were used for assessing its persuasive and motivational factors. The assessment is conducted with the aim to address the following questions:

How heuristic inspection could be done on motivational and persuasive elements situated in HCI?

What are the explanatory contributions of each grid in measuring the quality of Excerciser interfaces?


Figure 1. Screen captures of the exerciser.

Is there any ergonomic recommendations which can be learned to improve the interfaces?

The Exerciser was analyzed by the two authors who are HCI and e-learning experts. Their analysis was done separately and consensually combined in a second step.

Highlighted deficiencies with persuasive criteria In terms of credibility, the Exerciser’s interface inspires trust in the accuracy of its content. Explicit signs of its source and ownership (names and addresses, logo) and the school and professors recognized competence induces a sense of confidence in the learner. The system meets its promise: it contains exercises and solutions in college math.

On privacy, personal data and privacy are respected. A commitment is made early on by the site owners to ensure confidentiality and the user is reassured of the absence of risks that establishes a relation of trust.

In terms of personalization, only the user name (name, login, code) is explicit. We find no other signs of personalization (greetings, photo, etc.). The interface lacks features to promote personal closeness to the individual. It does not gradually collect the characteristics of the user allowing more personalized messages. Individualized information on the scores and results of the learners are known to the system (figure 2), but they are not used to in messages of support and motivation for the task. Individualization is therefore low. Yet, highlighting individualized features could raise user attention and interest in his/her training task,


ultimately fostering engagement. Furthermore, we note that there is no reference to group membership, nor to peer practices and performances and to tutor monitoring or other incentives that could induce some social pressure to engage in the task.

The attractiveness of the interface is weak. The aesthetics and appearance are relatively neglected, suggesting a somewhat cold and austere course, rather than the enthusiasm of an original exciting activity. Lack of attractiveness can deter the will to engage in the interaction. We note also that the design does not seek to be as close as possible to the values and needs of students. Finally, we note that the call to action is non-existent: the user is left to himself without a clear plan to follow or task sequencing incentives.

The Exerciser scores low over the dynamic factors. With regards to solicitation and commitment we note that besides the incentive mail to invite users to use the Exerciser (figure 1), the Exerciser does not sollicit, hunt for, or encourage learners. Suggestions and teasing that could awaken the curiosity are absent. The same is true for the sub-criterion of the initiation of the users who use the system once; they are not guided to undertake a first engaging action. Similarly, the incentives are reduced to the simple presentation of results to exercises (right or wrong), and it is the student himself declares success or failure (see figure 1).

With regards to commitment, we note the absence of a student engagement process in the design. The inspection of the interaction sequences does not reveal features to involve the individual in an gradual process that would lead to greater engagement towards learning goals. There are no systematic means to stimulate the user. The interface has no explicit feature to attempt to hold the student’s attention.

As commitment is not supported by the interface, we cannot find elements that would lead to any form of control or addiction of the user.

Highlighted deficiencies with motivational criteria The Exerciser gives full control to the learner, offering very little guidance apart from the navigation panel that indicates the topics and exercises covered.

The exercises represent a challenge in themselves, but the user can always choose to be given the answer instead of being required to find it by himself. This feature reinforces the control given to the user in his learning process, but at the cost of reduced challenge. However, the student can freely choose the exercises at the appropriate level of difficulty.

Learning is designed to be entirely independent, as the learner is not provided support or monitoring, neither from a human tutor, nor from colleagues.

Fantasy is absent altogether.

Confidence obviously varies considerably depending on student ability and personality, and the criteria should be considered from the perspective of whether the interface supports confidence for students of different profiles. The freedom to choose topics and obtain answers to exercises reduces the chances of feeling overwhelmed by exercises that are too challenging.

Sensory interest of the interface is very low and the comments on attractiveness of the persuasive interest apply here as well.

Cognitive interest also varies on a personal basis. From the perspective of HCI and e-learning application design, we find few means, if any, that can be leveraged to raise interest and that do not overlap with the other criteria.

Effort refers to a measure of use and cannot be derived from the interface itself, therefore it does not apply to a heuristic evaluation. However, from Lemieux and Desmarais (2013), we


have evidence that cognitive interest is generally low, with about half of 117 users spending less than an hour of total use. Yet, about 10 users spent over 10 hours and up to 140 hours, confirming that individual differences on the cognitive interest is highly variable.

Satisfaction is yet another factor that is assessed empirically instead of by inspection. We have no data to assess it for the Exerciser.

Discussion

We first note that some motivational factors of de Vincente et al. (2002) are not defined for an inspection process, namely effort and satisfaction. These can only be realistically assessed through experimentation.

However, given that we had data on the time spent by each student and the number of exercises completed, it must be emphasized that the array of levels of usage varies extensively, from a few minutes to over 100 hours. A reasonable interpretation is that some other factors are highly variable across individuals and is a determining factor that can overshadow all others. The most likely factors in are analysis are the cognitive interest or the challenge, and possibly a factor that is absent from deVincente et al. motivational criteria: the perceived value, which represents the importance that the learning goal represents to the user.

Another notable observation from the data is that about half of the users spent less than an hour in total, sometimes only a few minutes. Would this suggest that they do not have the same level of cognitive interest in mathematics or that they are not challenged by mathematics? Probably not since they have chosen to enroll in an engineering program that is itself challenging and involves mathematics. Can this be attributed to the perceived value of the Exerciser? Namely that they do not see the need to spend time in improving their math skills or the efficiency of doing so with the Exerciser. This is more likely, but it is not the sole interpretation.

Another reasonable interpretation lies in the persuasive factors. The low score on the attractiveness criterion and the quasi absence of means to support solicitation and commitment that would foster a progressive engagement is a sound explanation for the large proportion of users who spend a few minutes exploring the Exerciser without committing to its long term use. We could also refer to the fantasy and sensory interest as other means that could have been better deployed to address the low engagement.

Conclusion

This study aims to uncover the overlap between persuasive and motivational criteria, and the criteria that are unique to each, as well as their respective importance. The intent is that a better set of criteria can be obtained and used in an inspection process and help the design of an e-learning application.

We find that motivational criteria of fantasy and sensory interest can be considered similar to the attractiveness criterion of the persuasive grid. The persuasive grid has specific criteria such as credibility, privacy, and personalization that are specific. However, all of these criteria are not considered critical to explain the large variance in long term engagement that was observed for the usage of the Exerciser in practice.

Instead, motivational criteria such as cognitive interest, challenge and a criterion absent from deVincente et al. (2002), perceived value, are better candidate to explain the high engagement observed for some users. Conversely, the dynamic criteria of the persuasive grid are seen as


good candidates to explain why many users did go beyond cursory exploration of the Exerciser.

Finally, we observe that the data on long term usage of the application, which are indicators of the effort and possibly satisfaction criteria of the motivational factors, were critical in our analysis and cannot be obtained solely from expert inspection of an interface.

This study offers some guidance on what would represent a combined set of criteria to assess an elearning user interface, comprising both persuasive and motivational factors. Unfortunately, not all relevant motivational criteria can readily be assessed through the user interface inspection methodology and some require usage data, or questionnaire data.

Future work should focus on establishing the relative importance of the combined criteria. Experiments such as A/B schemes, where users are randomly chosen to be presented with one or the other user interface, could lead to interesting results in that respect.

References

Lemieux, F., Desmarais, M. C., & Robillard, P.-N. (2013). Motivation et analyse chronologique des traces d’un exerciseur pour l’auto-apprentissage. Sciences et Technologies de l’Information et de la Communication pour L’Education et la Formation (STICEF), (accepted).

Liu, C., Marchewka, J.T., Lu, J., and Yu, C.-S. (2004) “Beyond concern: a privacy—trust—behavioral intention model of electronic commerce”, Information & Management, pp. 127—142.

Nemery, A., Brangier, E., and Kopp, S. (2011). First validation of persuasive criteria for designing and evaluating the social influence of user interfaces: justification of a guideline In A. Marcus (Ed.): Design, User Experience, and Usability, LNCS 6770, pp.616-624.

Tràn, D. (2008). Motivation and persuasion in e-learning. Trends in E-Learning, 141, 2008. Retrieved from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.159.579rep=re

de Vicente, A., & Pain, H. (2002). Informing the detection of the students’ motivational state: an empirical study. In Intelligent tutoring systems pp. 933–943.


Specific Technologies for Persuasive Learning

The implementation of persuasive learning requires the development of novel technologies which

support the concepts of persuasive design. Bech has investigated the use of Augmented Reality as a

paradigm which shows the information right at the relevant location in the real world. Grund-

Sørensen has developed the Conceptual Pond for obtaining trends and user interaction. Sandborg-

Petersen has developed the EMDROS database system which is suitable for developing learning

objects which work with large corpus-based annotated texts.


Augmented Reality in Persuasive Learning using The Kaj Munk Case

Casper Justesen Bech

Department of Communication and Psychology, Aalborg University, Denmark

[email protected]

Through theoretical analysis, this paper aims at taking a close look at how immersion is achieved

when augmented reality is applied to a museum context. Augmented reality is by many regarded as a

technology with high value to the teaching and learning community. This is due to the possibility of

learners interacting with otherwise unavailable content and information. It is a central criterion of

success, for any medium, that the user experiences a sense of immersion. When augmented reality as a

medium is capable of changing user’s perception of reality, or in other words the world in which the

user perceives himself to exist within, the immersive part of the experience becomes central, as it is

the gateway to a mediated view of reality and thereby persuasion through augmentation. Through

chosen examples of immersive augmented reality usage within museum and learning contexts, the

theoretical implications of achieving immersion, while facilitating cultural heritage by persuasive

means, is discussed in relation to the possible use of augmented reality at the Kaj Munk Museum at

Vedersø Denmark. Keywords: augmented reality, immersion, persuasive design, persuasive immersion

Augmented reality and persuasive learning

Augmented reality as a technology has been defined in several different ways. The most

commonly used definition is also the earliest one by Azuma who states that augmented reality

systems must have the following three characteristics (Azuma, 1997):

Combines real and virtual

Interactive in real time

Registered in 3-D

As such the system is not defined by the technology used to combine the virtual content with

the real world environment, even though the display technology obviously has an effect on

how the augmented content is experienced. The ability to interact with the virtual content

must be in real time and this is central to the usage of augmented reality as a learning tool.

Real time simulation provides the ability to have direct interaction and exploration of

otherwise unavailable information and content through simulation. Research have pointed out

that this makes the technology ideal for discovery based learning (Angelopoulou et al., 2012),

which places the usage within a constructivist tradition of learning (Biggs et al., 2011). In

regards to how this relates to persuasive design, learning would for the purpose of this article

be seen as behavioral change. This is especially true when learning is seen within a museum

context, where the primary aim is to expand on the learners’ cultural understanding of a

certain subject or matter. In this specific case, which will be elaborated on later, the focus is

on the complex history and cultural significance of the Danish pastor and playwright Kaj

Munk (1898-1944).

If augmented reality is seen as a persuasive technology within this context, one could look at

how augmented reality as a medium persuades the learner to achieve the wished for outcome.

The ubiquitous nature of many augmented reality systems also makes the persuasive principle


of Kairos highly applicable (Fogg, 2003) (Mintz and Aagaard, 2012). The growing popularity

of smartphones, tablets and other powerful mobile devices, makes augmented reality systems

available in situ, when needed and therefore opens up for the possibility of combining

ubiquitous learning with augmented reality as a learning tool. The concept of Kairos can be

seen as more than simply timing an action to interfere at the opportune moment, in order to

achieve a certain outcome. A wider definition opens up for the possibility of Kairos (Gram-

Hansen et al., 2012) being a more central term in defining whether content is suitable in the

context where it is being presented. Not only how the content is presented is then in play, but

also what content is presented and when. In a museum context, this is highly important as

cultural heritage often is presented from a predefined perspective. How augmented reality is

experienced in terms of immersion within the system is then also related to Kairos and the

persuasive role of the augmented reality as a medium.

Virtuality and reality of an augmented possible world

Augmented reality has been described as being the middle ground between the completely

virtual, as in virtual reality and the completely real, as in the real world environment

surrounding the technology (Milgram and Kishino, 1994). Even the term augmented reality,

hints that the focus is to augment the perception of reality and thereby add to it. Augmented

reality is mainly made up of real world entities with virtual entities added to it, making the

virtual secondary to the primarily real world. Still, the virtually augmented content and the

way the virtual is combined with the real, is what makes the augmented reality experience. In

describing how immersion is achieved within augmented reality and how immersion effects a

successful persuasion, I will be looking into how these are affected by both the reality and

virtuality of augmented reality. It has been pointed out that a conceptualization of these two

concepts of augmented reality could be beneficial in relation to artistic potential and

application of augmented reality (Ma and Choi, 2006). This also relates to the idea of Kairos,

which is based on localized parameters that correlates to how the system is interactive. Then

again, interactivity further affects the way in which and to what degree immersion is

achieved.

For the purpose of analyzing how immersion has been achieved, I have chosen some

examples that share a context similar with that of the Kaj Munk Case. I will be using a model,

which takes the technological context into consideration, as well as how the virtuality and

reality of augmented reality affects immersion. The technological context sets the immersive

boundaries in regards to interaction and how the virtual content is being added to the real

world. The analysis will be based on the assumption of seeing the world as text, much like

how augmented reality can be seen as a mediator between the real world and a textual world.

Further on, it is important to note that augmented reality can be regarded as more of a media

aggregator, with the ability to augment the user’s perception of reality through all types of

media. Even so, this is highly dependable on the display technology. In theory the

augmentation of sight, which rarely stands alone, is only one possible augmentation.

The virtual content itself is regarded as text and at some level it is nothing but text. The basis

of all virtual content is computerized language, maybe the format be in letters or numbers.

The virtually added content will therefore be seen as text, even though it could be made up of

a wide variety of media. The content is delivered through the augmentation of the senses,

currently and primarily the sense of sight and hearing, but in theory senses as touch, taste and

smell could be mediated as well, thereby expanding further on the possible media used. For

the purpose on this theoretical analysis on immersion, the augmentation will be seen as text,

in which text is defined in widest possible way, so that everything that makes up the


augmented possible textual world is seen as text. Therefore, as augmented reality combines

the real with the virtual, the real world context will be considered as text as well.

The analysis seeks to evaluate on how the textual worlds are created in terms of the

technology used in relation to both the possible persuasive usage and how this relates to

immersion. As the prime characteristic of augmented reality is the mix of the real and the

virtual, the analyses will look at each of these aspects separately and then as to how the real

and virtual is combined. The persuasiveness and possibility of success in regards to the goal

and function of the system relate to how the combination of the real and virtual content

manages to immerse the user within the system. Different types of immersion apply in

different types of context. In this case, what often motivates a user to learn in a museum

context is how the system involves and excites the learner or user (Yuen et al., 2013). This

translates to the different types of immersion known from among others the gaming industry,

but Ryan (2001) also differentiates between different types of immersion, as will be shown

doing the analysis. One significant type of immersion, which also is the most significant in

terms of Kairos, is spatial and temporal immersion which covers the feeling of being

immersed in space and time. Per definition this type of immersion could often be seen as the

main focus of a museum, as it serves to give visitors or learners a feeling of how it must have

been to have actually lived in the setting on display. Now this does depend on the type of

museum. One type of could be being a museum of art. Here the focus is on the individual

pieces of artwork. Another type could be museums which focus on a certain period of time of

significance or a certain personality or place of significance. This last loosely defined type

would be the one covering the case concerning the Kaj Munk Museum at Vedersø.

These real world entities or the parameters of the real world context, define the technical

possibilities and the secondary entities, being the virtual augmentation. The real world

surroundings or the real world situational context plays a central role in terms of immersing

the user into the augmented reality world perception. When augmented reality is applied as a

learning tool within a museum context, they are almost always only locally usable at the

museum, meaning that they need the physical surroundings of the museum to function. In

combination, the augmentation creates a possible textual world and the type and degree of

immersion determine the possible persuasive outcome of the augmentation.

Immersion through augmentation

A commonly referenced example used when defining immersion and the immersive

experience, is that of how one can be absorbed into good literature. One can forget the

surrounding world and all focus centers around the literary possible world projected from the

text. It is important to note, that in theories of immersion, some have adopted Leibniz's

concept of possible worlds, with the important addition, that in the works of Ryan and others,

possible worlds serve as instruments of understanding rather than the Leibnizian idea of

possible worlds as something with a rather specific ontological status.

Different media and genres, with varying low or high tech formats possess different types and

degrees of needed immersion. Immersion is as such highly dependable both the mediated

context and the genre. Concerning augmented reality, which as previously stated, is an

aggregator of all types of media, the situational context defines the need for immersion and to

what degree. Of course the technology itself sets out limitations. As exemplified by the

quality of the display used, which relates to aesthetics and ultimately how real the augmented

content is experienced. Then again, the content does not always need to appear real in order to

be immersive or persuasive. This comes down to the situational context, as both fictional and

factional genres can immerse the reader. There is an underlying theoretical question of

whether immersion applies to augmented reality, in the same way as it does in regards to other


media. When one reads a book, plays a videogame or enters a virtual reality universe, one is

immersed into a possible world projected by the media. The basis of augmented reality is to

merge a possible world directly with the real physical world. As the user is not immersed into

a purely possible textual world, but rather a mix of the actual world augmented by the

possible textual world, immersion does not only define the transcendence between worlds, but

also the acceptance of a combined mixed reality view of the world. The traditional definition

of immersion as the suspension of disbelieves and acceptance of a presence in a possible

textual world is therefore challenged. The experience of presence within an augmented

possible textual world is untouched within augmented reality systems, as it is the added

virtual content, which still is the lesser of the two defining entities, that plays the defining

role. Embodiment of the user is still relevant, but there is no displacement of the perception of

presence. The possible textual world created by the augmented reality system still needs to be

accepted and the augmented content, again depending of its character, still needs to be

accepted, perhaps through suspension of believe but not necessarily. As such immersion, in

terms of augmented reality, has more to do with the acceptance of an augmented view of

reality, than the displacement of presence.

Persuasion

Finding the right persuasive strategy requires clear persuasive goals. In terms of cultural

heritage and learning within a museum context, I have found that even though there often is a

more or less specific general goal or purpose of the installation or museum. These goals tend

to be rather fuzzy. Often the sole object appears to be spreading information of a certain

subject, because of its historical importance. Therefore a persuasive augmented reality design,

would need clear persuasive goals beforehand and then approach these from a persuasive

design approach afterwards. When determining how immersion relates to persuasive power or

persuasive strategies, one could look at how a persuasive technology, as a persuasive media,

facilitates degrees and types of immersion. If an augmented reality system provides metadata

for installments at a museum expedition and relates these to other related pieces at the

museum. Then these could be seen as a way to tunnel the user through a predetermined route.

Here the immersive experience stands in the background, as the tunneling of the user is a

relatively simple persuasive tool (Fogg, 2003), which does not take on a significant role in

terms of attention. Other types of persuasive media roles could be seen as facilitating more

significant degrees and types of immersion. This establishes a relation between the role of the

persuasive media and degrees and possible types of immersion.

Augmented Immersion and Persuasive Immersion

In terms of immersion, the relationship between the physical real (reality) and the

augmentation (the part of reality that is assigned as such) can be described in the following

manner. Reality itself, as presented by the curators of an exhibition, can obviously be the

focal point of experience. An augmentation understood as a caption or meta-text can also be

the focal point of experience. In the case of augmented reality, that is, the combination of the

two, the presentation can be such that reality is favored and the augmentation remains in the

background of attention. Conversely, the augmentation can also be favored over reality such

that it becomes the primary object of attention. This constitutes an ascending order of

augmented inversion (as illustrated in Figure 1).


Figure 1: The relation between the strength of augmented immersion and strength of

persuasive immersion.

In Fogg’s thinking of persuasive principles (Fogg, 2003), three main categories are

established. Interactive systems can function as tools that make desired actions easier to do.

Interactive systems can also function as social agents that create relationships by praising and

rewarding the user. Finally, interactive systems can serve as simulations, offering the user

experiences and room for active experimentation. This division is known as the functional

triad, and is understood as aspects of persuasion that complement each other. They differ,

however, in how they facilitate types of immersion. Under the tool perspective, the focus of

attention remains locked on the subject being presented because tools mainly are used to

accommodate the user’s interaction without drawing much attention to itself. In the agent

perspective, immersion may become more pronounced because of the more direct and

personified interventions. And finally, the simulation aspect allows for stronger aspects of

immersion because attention is drawn directly to the dynamic and interactive presentation of

material. When in this fashion, a user is allowed to ‘play around’ and explore consequences of

actions, the persuasion takes on a stronger form of immersive nature.

Examples of augmented possible textual worlds

In the 2011 study “Using augmented reality and knowledge-building scaffolds to improve

learning in a science museum” an augmented reality system at a science museum in used to

show students how the human body can conduct electricity (Yoon et al., 2012). The research

design focusses on how digital augmentation can function as a scaffolding technique together

with collaboration, written and verbal instruction etc. In regards to immersion, the study did

not focus on this, even though it clearly was an underlying aspect of getting the students to

achieve cognitive gains through the augmented reality experience. The real world situational

context here was a science museum, in which students in groups were introduced and

presented with tasks that included using the augmented reality system, in different forms, to

answers questions. It is important to note that the augmented reality system used in this

specific case was stationary, thereby eliminating much of the informal discovery based

learning that was mentioned earlier. But what this study clearly shows is that the level of

interaction plays a large part in creating an immersive system. When the students managed to

conclude the circuit with their hands, the digital augmentation would be activated to illustrate

the flow of electricity. The simplistic nature of this example shows that augmented reality can

be used in a learning environment to build knowledge. Here augmented reality not only serves

as a theatrical prop to make the museum more interesting, but also as a knowledge building


scaffold, that develops cognitive gains. The term cognitive immersion (Bjork and Holopainen,

2005) can be applied here. Through problem solving, although not at a complex or advanced

level, the students apply reasoning and thereby engage themselves cognitively with the

system. This leads to an immersive experience and acceptance of the augmented perception of

reality. The physical involvement in completing circuit, the interactivity with the system

which is one of the key characteristics of augmented reality technology, further immerses the

students into an augmented view of reality. According to Ryan, the mediating of the body,

thereby using the body as an interface with the system, leads to a more natural interaction,

thereby strengthening the immersive experience (Ryan, 2001).

It is important to note, that the immersion into a mixed reality world where projected

animations run over once arms to illustrate current, is only accepted as part of a possible

textual world, if it is believable. In this context believable means realistic or artistically

pleasing. According to Coleridge this is caused by the “willing suspension of disbelief”. Here

the user of the system chooses to believe that the fictional augmentation is real, in order to be

immersed into the possible world projected by the system. The learners are accepting, what

could be seen as a fictional augmented view of reality, but in accordance with literary theory,

a possible world, in which the augmented content is possible. The transparency of the medium

(Ryan, 1994), the projection of the virtual content into the real world, makes the content

appear even more real. This is caused by the user not having to look at it, through a screen.

The computer generating the virtual content is not visible or at least it appears as if it is not

visible. Different systems offer different degrees of transparent computing, which is closely

related to the idea of ubiquitous computing. As such the baseline of possible immersion is

highly dependable on the technology used. Yuen, Yaoyuneyong and Johnson differentiate

between three technologically dependable levels of immersion (Yuen et al., 2013). The way in

which the seamlessness of the virtual content is blended into reality is here the main defining

factor in which the display technologies are categorized as either low immersion, high

immersion or total immersion. Video projectors, as the ones used in the augmentation

mentioned above are categorized as “low immersion”. This is due to the immobility and

psychical visibility of the display technology being used. As an example, when using a

projector as display technology, one can actually see the light that the projector makes, a

byproduct of the augmentation, rendering it less immersive and this light is not a part of the

possible world the augmentation creates. At the other end of the scale, technologies such as

HMDs or display implants are categorized as “total immersion”, as these in theory offer

higher mobility and a high level of seamlessly integration between the real and virtual

content. The type of immersion referenced to here is spatial immersion. Spatial immersion, in

the terminology of Yuen, Yaoyuneyong and Johnson, means the experience of being

surrounded by aesthetically and enjoyable augmented content. This clearly applies in the “be

the path” augmentation as well. If the augmented content is not clearly understandable or

believable, then there is a chance that the system fails to immerse the users with the possible

world and thereby failing as a learning tool. The setup discussed here is rather simplistic and

spatial immersion does not play a significant role, as the augmented content is rather limited,

which means that the possible world is limited as well. The augmentation could here be seen

as a social agent, in the way that cooperation between the students is needed in order to

complete the circuit. The cognitive immersion is thereby established through the persuasive

role of agent.

Spatial immersion plays a more significant role in the augmented reality system used at

Sutton Hoo archaeological site near Woodbridge England. In the 2012 paper “Mobile

Augmented Reality for Cultural Heritage” a augmented reality system is utilized to engage

visitors at the site by augmenting content to provide information and gamify the experience.


Several studies have showed how augmented reality can be used in a learning context, to

engage the learners (Yoon et al., 2012). In terms of immersion, this study is interesting for

two reasons. The first being uses highly mobile devices as display technology for the

augmented reality system, making the system more ubiquitous or agile and giving the user a

higher degree of interaction – both of which means a greater degree of immersion in terms of

technically ability . Secondly, the learning context shares some basic premises with the Kaj

Munk Case. The system relates relevant information and interpretation, at times when these

would appear relevant for user, based on positioning and user input. This directly relates to

the idea of Kairos, in which persuasive tools as tailoring and suggestion can be seen as

applied in order to provide the user with the right information at the opportune moment.

Further on, there is a social element, in how it is possible to cooperate with museum visitors

at a distant location. Again, the system takes on the persuasive role of a social agent, as it

facilitates cooperation. This social interaction could be seen as facilitating emotional

immersion, as the visitors at each museum becomes dependent on each other in order to

complete the goals set up by the game established within the possible textual world.

The Kaj Munk Case

At concept level, there have been ideas of using tablets to augment the visiting experience at

the Kaj Munk Museum. The aim is to establish a textual possible world, that serves to

persuade the user as to fulfill the museums wishes for better spreading knowledge and maybe

acknowledgement of the life and significance of Kaj Munk. One of these ideas involves

bringing Kaj Munk back to life through the augmentation of an actor playing the role as the

pastor and playwright. This would clearly bring the augmentation in the foreground of

attention. In terms of the persuasive role of such a possible textual role that actually simulates

the main character of the textual role at Vedersø, this could provide several interesting types

of immersion of varying degrees. In terms of spatial and temporal immersion, it is clear that

the combination of not only being in the real physical home of the possible textual worlds

deceased main character – but also experience him in character, provides a high level of

immersion and persuasive possibility. Emotional immersion is clearly involved as well, but

this is of cause related to the visitor’s relation to Denmark and national history. Combined

with Kairos, the experience could be customized to fit the visitors need through user input.

Emotional immersion could then potentially prove to make the experience especially

persuasive – although there are some clear ethical implications related to this, which lies

outside of the scope of this paper.

Another idea is that the collected digitalized works of Kaj Munk, from the Kaj Munk

Research Center at Aalborg University, could be utilized through relevant hyperlinking

between relevant locations and or objects at the museum in Vedersø. This could serve as

cognitive immersion, by providing the material at the right moment, in order to trigger a

relation between the real world environment at Vedersø and the augmented digitalized works

of Kaj Munk.

Furthermore, as groups often visit the museum together,it would be foolish not to utilize this

in terms of persuasive immersion. The augmentation could act as a social agent, in the sense

that the augmented possible textual world could motivate debate and discussion within the

group of visitors. Thought provoking examples of Kaj Munks writing could be augmented, at

the right moment and in correlation with the group of visitors’ physical placement at the

museum. This could lead to a greater degree of cognitive immersion among other types of

immersion. If the augmented reality system takes on several persuasive roles, being both the

role of persuasive tools, that of the social agent and at last but not least – the role of


simulation. Through the role of a social agent, the system could try and provoke the visitors to

discuss some of the more controversial statements given by Munk. This could lead to a

greater degree of cognitive and emotional immersion. If the system takes on role of

simulation, it could facilitate a possible textual universe, in which it simulates what could

have happened, had Kaj Munk not been murdered by the Gestapo in 1944. This type of

persuasive role, especially if the system allows the visitors to have choices and then play out

alternative World War II scenarios, could possibly mean an even greater degree of cognitive

and emotional immersion. This could be beneficial for visitors and further strengthen the

chance of visitors leaving with a greater knowledge of Kaj Munk and his historical

importance.

References

Angelopoulou, A., Economou, D., Bouki, V., Psarrou, A., Jin, L., Pritchard, C., Kolyda, F.,

(2012). Mobile Augmented Reality for Cultural Heritage, in: Venkatasubramanian, N.,

Getov, V., Steglich, S. (Eds.), Mobile Wireless Middleware, Operating Systems, and

Applications, Lecture Notes of the Institute for Computer Sciences, Social Informatics

and Telecommunications Engineering. Springer Berlin Heidelberg, pp. 15–22.

Azuma, R.T., (1997). A survey of augmented reality. Presence-Teleoperators Virtual Environ.

6, 355–385.

Biggs, J.B., Tang, C.S., Society for Research into Higher Education, (2011). Teaching for

quality learning at university. Open University Press, Maidenhead :McGraw-Hill.

Bjork, S., Holopainen, J., 2005. Patterns in Game Design. Cengage Learning.

Fogg, B.J., (2003). Persuasive Technology: Using Computers to Change What We Think and

Do. Morgan Kaufmann.

Gram-Hansen, S.B., Schärfe, H., Dinesen, J.V., (2012). Plotting to Persuade – Exploring the

Theoretical Cross Field between Persuasion and Learning, in: Bang, M., Ragnemalm,

E.L. (Eds.), Persuasive Technology. Design for Health and Safety, Lecture Notes in

Computer Science. Springer Berlin Heidelberg, pp. 262–267.

Ma, J.Y., Choi, J.S., (2006). Augmented Reality as Perceptual Reality, in: Zha, H., Pan, Z.,

Thwaites, H., Addison, A.C., Forte, M. (Eds.), Interactive Technologies and

Sociotechnical Systems, Lecture Notes in Computer Science. Springer Berlin

Heidelberg, pp. 7–10.

Milgram, P., Kishino, F., (1994). A TAXONOMY OF MIXED REALITY VISUAL DISPLAYS.

IEICE Trans. Inf. Syst. Vol E77-D, 1994.

Mintz, J., Aagaard, M., (2012). The application of persuasive technology to educational

settings. Educ. Technol. Res. Dev. 60, 483–499.

Ryan, M.-L., (1994). Immersion vs. Interactivity: Virtual Reality and Literary Theory 1994.

Ryan, M.-L., (2001). Narrative as Virtual Reality: Immersion and Interactivity in Literature

and Electronic Media. Johns Hopkins University Press.

Yoon, S.A., Elinich, K., Wang, J., Steinmeier, C., Tucker, S., (2012). Using augmented

reality and knowledge-building scaffolds to improve learning in a science museum.

Int. J. Comput.-Support. Collab. Learn. 7, 519–541.

Yuen, S.C.-Y., Yaoyuneyong, G., Johnson, E., (2013). Augmented Reality and Education:

Applications and Potentials, in: Kinshuk, Huang, R., Spector, J.M. (Eds.), Reshaping

Learning. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 385–414.


Intuitive Surveying & Quantification of Qualitative Input

through the Conceptual Pond Christian Grund Sørensen

Center for Computer Mediated Epistemology, Aalborg University, Denmark

[email protected]

“The Conceptual Pond” is an application for assessment designed to gather qualitative

input through an intuitive, visual interface. The process of using the application serves as

a conceptual aid for personal reflection as well as providing an evaluation system with the

ability to transform this input into quantitative data. The application thus serves three

distinct yet parallel purposes:

1. One of assessing and evaluating through an interface designed with a focus on

intuitive handling, visual language and facilitating input of a qualitative nature.

2. One of processing such assessment of qualitative input so it is transformed into

quantitative, comparable representations.

3. One of supporting learning, reflection and immersion through agency in a digital

environment.

The Conceptual Pond is developed as a response to challenges of evaluation related to the

EuroPlot-project (an EU-supported research project under EACEA), in particular the Kaj

Munk Case, focusing on the dissemination of biography, works, and impact history of

Danish vicar, journalist, and playwright Kaj Munk. The application is designed

implementing persuasive principles (Fogg, Oinas-Kukkonen et al.), learning taxonomies

(Biggs & Tang, Bloom) together with a rhetorical understanding of kairos.

This paper includes the evaluation of several use cases as well as two modifications to the

original application design supporting pre-literate children as well as implementing a

temporal perspective to application functionality.

Keywords: Conceptual Pond, persuasive technology, assessment, intuitive

Introduction

In Western societies (and most other societies) it is the prevailing order that most learning is

done within the framework of an organized school system. Teaching a curriculum through

various methods shapes the knowledge and reflection of pupils and students. Progress in the

field of reaching intended learning outcomes are usually monitored through exams, tests, and

assessments. (Biggs & Tang, Bloom) Despite of changes in pedagogical method and learning

designs the dichotomy and interaction between learning and assessment is stable. Even if

grading is not an issue the mere evaluation of learning is necessary for the planning of future

learning events by the educator.

In a number of contexts, however, formal evaluation in the shape of exams and tests are a less

viable option. Much learning is done, insights gained and reflection facilitated in a number of

less patterned learning environments (Falk, Hooper-Greenville). In the area of cultural

mediation learning is only partly consistent with the formal education system. It is often based

on casual interest, fragile preceding knowledge and fluctuating attention. So, in cultural

mediation or pedagogical work with small children for example, another approach to learning

and assessment may be appropriate.


Development of The Conceptual Pond has been made in order to facilitate the gathering op

input of a complex nature such as impressions, reflection, opinions, and less categorized

knowledge. These are gathered through a simple interface with an emphasis on overview,

intuition and interface tangibility.

At the same time this diverse input is compiled in the application and is transformed into

generic data suitable for data presentation and cross tabulation thereby transforming

qualitative data into material suitable for quantitative analysis.

The intention has been to develop an assessment tool shaped to fit the interests and

expressions of the users rather than fitting these interests into too firmly predefined

frameworks. User adaptation is supported by the interactivity of the digital format.

This paper serves to present a pilot study of The Conceptual Pond and the theoretical

framework behind it as well as suggesting areas for further research.

The Conceptual Pond is designed to solve tasks primarily in the field of learning, social

sciences and human sciences. The theoretical background and functionality however makes it

no less relevant in other sectors such as human computer interaction, psychology, or social

sciences. The need for analyzing humanly expressed thoughts is central in a broad variety of

academic, pedagogic, cultural or societal projects. Use cases referred to in this paper should

demonstrate this applicability in several environments.

Background

The Conceptual Pond is originally developed to meet challenges of evaluation in the

environment of the EuroPlot-project (an EU-supported research project under EACEA). Part

of this project has a focus on the provision and dissemination of cultural content. The Kaj

Munk Case has been explored by Aalborg University and the Kaj Munk Research Center. Kaj

Munk (1896-1944) was a leading Danish theologian, playwright, journalist and writer. His

works comment the political and philosophical issues of his era as well as discussing

existential and religious subject of the modern paradigm. Kaj Munk also contributed

substantially to Danish WWII history by being executed by German authorities in 1944 for

his nonviolent resistance to the German occupation. Ideas from Kaj Munk were inspirational

for the implementation of truth committees in the South African reconciliation process after

apartheid, thus proving the heritage of Munk to be relevant in a history of impact.

Due to the complexity of the content of this cultural mediation several approaches have been

made providing digital learning material content (Hansen, Hansen, Sørensen, Øhrstrøm 2013).

Archival access is facilitated through The Munk Study Edition (Petersen & Øhrstrøm)

offering fully searchable annotated texts of Munk’s plays and sermons. The application is

based on an EMDROS-database (Petersen 2004) The Munk Study Edition is supplemented

with GLO’s (Generative Learning Objects) (Hansen, Hansen, Sørensen, Øhrstrøm 2013).

Working with these multimodal learning processes raised the attention to the need of

acquiring valid feedback of not only knowledge but also interpretation and impression. For

this reason The Conceptual Pond was designed in cooperation with Mathias Grund Sørensen

(Sørensen & Sørensen).

Persuasive assessment

As mentioned in the introductory remarks assessment is a consistent part of most common theories of learning. In the neoclassic understanding of cognitive domain in Blooms Taxonomy (Bloom) assessment is a vital part. In the Structure of Observed Learning Outcomes Taxonomy of Biggs (Biggs & Tang) evaluation and assessment play an equally important role – emphasizing the importance of


observation in the term “Observed Learning Outcomes” of the SOLO taxonomy. In the Anglo-Saxon learning paradigm to which the two learning theories belong assessment plays a more dominant role than in Scandinavian school tradition. However, since evaluation is an integrated part of contemporary educational practice. It is necessary also for the EUROPlot-project to develop and present suitable assessment tools for the evaluation of the individual student/users as well as groups. Nevertheless, as Odendahl points out: “Abstract principles have to be translated into real tests for real students … ”How” decisions include question format: selected-response tasks, in which the student selects from among answers that are provided on the test, or constructed-response tasks, in which the student generates a product or performance.” (Odendahl, 2011, p. 8). The process is in casu assessment related to the Kaj Munk Case of a two-step nature (Rossman & Rallis): 1. The gathering of valid qualitative input. 2. The adaptation of these data into an environment suitable for further analysis. The first part of this two-step model implies the motivation and persuasion of the learner to take part in the process of assessment. In a regular school system at any level this may often be a minor challenge. Moving into the area of cultural mediation users usually feel less obligation to take part in surveys etc. This calls for another motivational approach than the Benthamian concept of the omnifocal Panopticon in which prisoners were subject to assessment in the shape of surveillance. The Panopticon is used also as a societal metaphor by Michel Foucault (Foucault, 1977; Jespersen et al. 2007). Perhaps a third aspect needs to be added to the Rossman & Rallis approach: 3: The implementation of qualitative input into a persuasive process for enhancing reflection. Tools and principles for meeting this challenge may be found in discourse in the field of persuasive technology. The basic discourse in the persuasive technology field is rooted in the research of behavioral psychologist B.J. Fogg. In his book “Persuasive Technology – Using Computers to Change what we Say and Do” he combines observations in the human computer interaction field with a psychological understanding offering a number of methods or devices to persuade the user into adapting a certain attitude or behavior “without coercion or deception” (Fogg, 2003). Many of the “classic” persuasive devices of the 2003 book require a certain kind of monitoring, assessment or feedback to function as intended. This also applies for Foggs’ persuasive tools, namely tailoring, conditioning, reduction (to some extent), and in some cases also tunneling, and suggestion. These persuasive devices most often function with the assistance of interactive feedback that allows for proactive agency decided by the feedback of past choice, such as product recommendation systems utilizing past user purchases.

Figure 1. Assessment in Persuasive technology

Device: Surveillance Peer comparison Self-monitoring

Example:Panopticon CCTV Exams Endomondo (1) Educational games Endomondo (2)

Assessment and evaluation is often at the center of technologies related to self-monitoring and surveillance (Oinas-Kukkonen). In Figure 1. the role of assessment in Persuasive Technology is defined as an axis with three elements: Surveillance, peer comparison, and self-monitoring. A not very benevolent interpretation could be to make a parallel between surveillance and assessment in the shape of exams and tests. This may be partly true. At the same time peer comparison is an important factor in educational assessment systems, and even the element of self-monitoring is vital part of the motivational process for the individual student. The pinpointing of exams to one single point at the axis is therefore reductionist. The examples may be pinpointed differently on the surveillance - self-monitoring axis depending on the actual use. Therefore


Endomondo, the sports tracker, is also pinpointed as (1) or (2) depending on what features are used in the app. If the tracker is applied only for personal individual use it supports and motivates through self-monitoring. Once results are shared on social media a more or less voluntary submission to surveillance is accepted. In the Persuasive Technology field such dynamics are heavily discussed, especially in the areas of health-care and prevention. A classic example is Fogg’s heart-rate monitor (Fogg). In an extended form it plays a role in studies of persuasive motivation (Munson, Lauterbach, Newman, and Resnick) where the role of assessment is linked with a social network site in order to benefit from the persuasive power of social comparison and peer support. Another example is the studies of weight loss websites by Lehto and Oinas-Kukkonen (Lohto & Oinas-Kukkonen)). They underscore the worth of assessment in the context of social facilitation, which is closely related to self-monitoring and social comparison.

Kairos

In several aspects the ancient rhetorical concepts are helpful in designing with persuasive

intent (Hasle) Bringing the concept of kairos into play in a postmodern paradigm facilitates a

three-dimensional view to the potential of The Conceptual Pond. Reaching kairos is the art of

designing and applying technology at:

1. The opportune time

2. The opportune place (locationary or virtual)

3. The opportune manner (also medium)

This suggests an ecological approach focusing on adaptable systems design suitable for any

learning environment.The Conceptual Pond facilitates the implementation of all three aspects

of kairos.

From a perspective of reflection it is imperative that some acts of evaluation are carried out

immediately after receiving the impression. In the use case of the teenagers evaluation of

impressions from viewing the Kaj Munk play, they still had the experience in vivid memory

when accessing the survey. Since TCP is part of technology enhanced learning and is

available at all kind of online devices it may be used at any given time. Another aspect of

kairos is the time elapsing carrying out a task. In relation to this it is promising that tests

displayed response times between 95 and 140 sec. on average which appears considerably less

time consuming than comparable questionnaires.

Since learning and reflection takes place in a number of environments it is imperative that the

evaluation system may be available everywhere. This is the case since The Conceptual Pond

is itself everyware in the sense that technology allows for it being used even in the open

country by the Kaj Munk Vicarage Museum, at a kiosk at the library or at a school computer.

The third aspect of kairos refers to the function and characteristics of the system itself. Does it

offer content in a contextually relevant way? In relation to The Conceptual Pond the question

primarily is one of mediation. As shown in the use cases the choice of technology and mode

proves to be convincing. Apart from the few motorical problems handling a non-touch device

the system seems fully sustainable. Using The Conceptual Pond should facilitate reflection

and assessment in a helpful way thus mediating in the opportune manner.

The Conceptual Pond

On the basis of these insights from the areas of learning and Persuasive Technology the conceptual Pond is designed as an intuitive, visual interface for assessment. Depending on use TCP facilitates a type of qualitative assessment which may be relevant at any point of the axis of Figure 1. On the other hand it is obvious that the complexity of the possible input from


users makes it less helpful for the rather more simplistic answers that will usually be collected in a multiple choice test. The Conceptual Pond is designed with a simple graphical interface. This is important, as the application should be intuitive and easy to navigate even for primary school children and the technologically challenged. Navigation is a simple drag-and-drop movement and the user is always able to alter her choice and change position of the marker. The user has the command of a number of markers, each of them marked with a word or a short sentence. The user can insert non-pre-designed words of her choice thus extending her expression (if this is not prohibited in the evaluation set-up). Unused markers stay in the right side of the screen, unless the user chooses to leave them otherwise. The markers are clearly marked as moveable boxes and have a clear inscription of the corresponding word. In special situations the markers could be substituted with pictures, colors or symbols as previously mentioned. The left and center part of the assessment screen is dominated by a circle of relevance. This circle of relevance marks the area in which the words relevant to the user, describing her denotations and connotations related to the subject, are dragged and dropped at will. The center of the circle of relevance is a marked in a clear color fading slowly into a lighter color. The position relative to the center signals the strength and importance of the expression. Using gradients of the colors green enhances the intuitive impact of the interface. As noted in research (Ham & Midden) the colour green is intuitively recognized as something positive and thereby reduces the cognitive load of the user.

Figure 2. The Conceptual Pond (simplified)

Figure 2. shows a simplified example of the use of The Conceptual Pond related to the Kaj

Munk Case. Results of actual use case presented below.

Use cases

The Conceptual Pond has been implemented at a number of tests. One of these tests (Ex 1:

Reception of Munk’s play “The Word”The Word) is presented in detail.


Table 1. The Conceptual Pond full-scale testing

Ex Period Number Participants Place Assessment

1 October 2012 23 Confirmation class

13-14 y.o.

Church House

Nørager

Reception of Munk’s

play “The Word”

2 November 2013 25 Library users

8 – 76 y.o.

Mun. library

Nørager

Satisfaction and

interests at library

3 March 2013 40 Philosophy

Students

Aalborg

University

Reflections on Munk’s

play “Love”

4 April 2013 29 Various adults

attending Easter on

Facebook

Internet survey

Denmark

Reception and

reflection on Easter on

Facebook-project

5 June 2013 83 Various adults

attending Danish

Church Days

Aalborg

Conference

Center

Opinions on important

subjects in Church

communication

6 2013

planned

25

approx

Preschool children Kindergarten Testing TCP featuring

images instead of text

Kaj Munk: The Word

A group of 23 teenagers (13-14 years old) viewed a selected scene of the play “The Word” by

Kaj Munk (1962) in a well-known 1955 movie adaptation of Carl Th. Dreyer. The movie was

rewarded the Golden Lion Award at the 1955 Venice Film Festival. Despite an antiquated

visual language featuring slow dialogue and black/white aesthetics the scene of the

resurrection of a woman deceased in childbirth proves to be very moving and is an iconic

moment in Danish dramatic tradition.

Question was, however, how would this group of teenagers react to the screening? What are

their impressions and reflections? At the same time it was vital not to influence the

participants in the survey. Immediately after the screening the teenagers were given a very

brief introduction to The Conceptual Pond (2 minutes). Remaining in silence they accessed

the system individually contributing with their observations and reflections. The interface was

accessed on a laptop PC similar to devices used at school thus eliminating cognitive stress in

operation. The laptop was placed in a kiosk setup with the screen turned away in order to

secure privacy of the users and avoid possible peer pressure that could bias the results. Each

of the 23 teenagers initiated their session filling out a few pieces of generic information

(name, gender, age). Subsequently all contributors accessed the interface and reported their

impressions to the system.

This process was monitored from some distance to respect privacy and at the same time

record the time lapse. Average operating time was 105 seconds ranging from 25 seconds to

180 seconds. All participants except 2 performed between 75 and 135 seconds. Only one user

displayed any hesitation in using the system. None required additional instruction. All 23

participants explicitly preferred The Conceptual Pond to a paper questionnaire. 9 users had

specific comments. 2 suggested minor changes in design. 7 applauded the user freedom

compared to other evaluation methods (such as questionnaires and qualitative interviews).


Figure 3: Graphic presentation of case results “The Word”

Results of the results of the teenagers were transformed from quantitative expressions to

qualitative data. Production of graphic was done instantly in TCP-application.

Development: Implementing temporal sequence

In a number of contexts a temporal sequence assessment may prove useful. The linking of

several proceeding psychological tests may be an example (Sørensen & Sørensen).

Development in psychological evolution may be stored and analysed through the system.

In a dynamic understanding of kairos this would entail the opportune moment turning into

several opportune moments linking together in kairos.

Development: Adaption to the illiterate

Recognizing the needs of children and other illiterate people markers may be exchanged by

images or pictograms. In a pedagogical analysis situation as an example markers describing

everyday situations with words may be exchanged by pictograms. Placing

the pictogram into the circle of relevance at a desired position may be a

valid expression of trust or anxiety in a given situation.

Implementing a visual language may also benefit literate users of certain

cognitive styles (Riding & Rainer).

Conclusion

Applying the Conceptual Pond to appox. 200 people in quite different contexts with quite

good results and good feedback allows the conclusion that The Conceptual Pond is a helpful

tool. It facilitates easy and intuitive gathering of – often complex – qualitative information. It

facilitates easy quantification and instant access to this data as well as making I possible for

further cross tabulation.

The greatest experience working with The Conceptual Pond has actually been the expressions

of the users in relief that the design is intuitive and not another complex questionnaire-based

survey. Recognizing the number of agents using questionnaires it is not surprising that a

certain fatigue builds up which may also damage the acquisition of valid data.

Nevertheless a the development and refinement of the conceptual

Further research


Testing so far of The Conceptual Pond has pointed to it being persuasive, easy to use, and

helpful in the acquisition of qualitative data. Nevertheless TCP needs to be tested further

refining the interface design to support clear user expression and secure assessment does not

facilitate any kind of bias.

Avery promising field is the area of tailoring TCP for certain assessment goals.

Another promising field is the implementation of a temporal flow in assessment, as well as

creating a fully functional interface for illiterates, children and the handicapped.

References

Behringer, R., Gram-Hansen, S.B., Soosay, M., Mikukecka, J., Smith. C., Winther-Nielsen, N,. and Herber, E.,

(2013) Persuasive Technology for Learning in Business Context. ASCENT 2013 Conference, Kuala Lumpur,

11.-12.April 2013.

Biggs, J & Tang, C. (2007) Teaching for Quality Learning at University (3rd edn) Buckingham: SRHE and

Open University Press

Bloom, B.S.(1956) Taxonomy of Educational Objectives, the classification of educational goals – Handbook I:

Cognitive Domain , New York

Fogg, BJ (2003) Persuasive Technology - Using Computers to Change What We Think and Do. 2003, San

Francisco: Morgan Kaufmann Publishers

GLOMAKER: http://www.glomaker.org/ (2013) Gram-Hansen, L. B. (2009) “Geocaching in a persuasive perspective”, in “Proceedings of the 4th

International Conference on Persuasive Technology”, ACM New York 2009

Gram-Hansen, S.B., Scharfe, H, Dinesen;J (2011) D.3,1 Persuasive Learning Design Framework 7/2011

Europlot deliverable

Gram Hansen, S.B., Gram Hansen, L.B., Sørensen C.Grund,, Øhrstrøm, P Deliverable 5.3 PLOTLearner v3: The

Kaj Munk Case

Ham, J., & Midden, C. (2010). Ambient persuasive technology Needs little Cognitive Effort: The Differential

Effects of Cognitive Load on Lighting Feedback Versus Factual Feedback. Lecture Notes in Computer Science,

6137/2010.

Hasle.P., Christensen, A.K.K (2007): Classical Rhetoric and a Limit to Persuasion. Persuasive 2007, LNCS

4744. Springer

Hasle, P. (2006) The Persuasive Expansion - Rhetoric, Information Architecture, and Conceptual Structure, in

ICCS 2006, H. Schärfe, P. Hitzler, and P. Øhrstrøm, Editors. 2006, Springer

Lehto, T., & Oinas-Kukkonen, H. (2010). Persuasive Features in Six Weight Loss Websites: A Qualitative

Evaluation. Lecture Notes in Computer Science , 6137/2010.

Munson, S., Lauterbach, D., Newman, M., & Resnick, P. (2010). Happier Together: Integrating a Wellness

Application into a Social Network Site. Lecture Notes in Computer Science , 6137/2010.

Odendahl, N.V. (2011). Testwise - Understanding Educational Assessment, Volume 1, Lanham 2011.

Riding, R. & Rayner, S (2000) Cognitive Styles and Learning Strategies., London

Petersen, Ulrik (2004). Emdros - A Text Database Engine for Analyzed or Annotated Text. In: ACL, COLING

2004 Geneva, 20th International Conference on Computational Linguistics, Volume II. Proceedings, pp. 1190-

1193.

Sandborg Petersen, U. & Øhrstrøm, P. (2010) Det Virtuelle Forskerværksted for Kaj Munk Studier in

Munkiana, 44/2010

Grund Sorensen, C., (2013) Easter on Twitter - a Digital Pilgrimage on the Mobile in “Viva Vox Evangelii”,

Evangelische Verlagsanstalt. Leipzig

Grund Sorensen, C. & Grund Sørensen, M. (2013) The Conceptual Pond – a Persuasive Tool for Quantifiable

Qualitative Assessment in “Emerging Research and Trends in Interactivity and the Human-Computer

Interface”, IGI Global

Rossman, G.B. & Rallis, S.F. (2012) Learning in the Field. SAGE, Thousand Oaks 2012.

Tørning K., Oinas-Kukkonen H. (2009) Persuasive System Design: State of Art and Future Directions. ACM

International Conference Proceeding Series, Vol. 350,


Architecture of applications built on Emdros:Case studies in systems for persuasive learning

Ulrik Sandborg-PetersenDepartment of Communication and Psychology, Aalborg University

Nyhavnsgade 14, 9000 Aalborg, [email protected]

This paper describes the nature and scope of Emdros, a text database engine for annotated text. Three case-studies of persuasive learning systems using Emdros as an important architectural component are described, and their status as to participation in the three legs of B.J. Fogg's Functional Triad of Persuasive Design is assessed. Emdros is shown to be able to form the foundation for any persuasive learning system whose primary function involves text being stored in or retrieved from a database. Similarly, Emdros is shown to be able to provide important support for all three legs of the Functional Triad.

Keywords: Software Architecture, Persuasive Learning Systems, Text Database Engines

Introduction

The EuroPLOT project aims to make it easy for educators to create, share, and reuse “Persuasive Learning Objects”. This goal has been attained so far in a number of systems, including the so-called “PLOTLearner” text-driven language learning tool, and the “Kaj Munk Study Edition”, which is a software program for the display and study of the works of a Danish poet, pastor, and playwright, Kaj Munk. In this article, we explore the architecture of each of these systems, and also describe the architecture of a third system, namely the “Kaj Munk Virtual Research Laboratory”. All three of these systems are persuasive in nature, and all three of them have a particular database management system as their architectural underpinning, namely the Emdros database engine.

The rest of the paper is laid out as follows. We first describe Emdros, its nature and its scope of use. We then describe the three systems under case-study, including their architectures, and we show how each of them embodies some aspects of the Functional Triad of B.J. Fogg [Fogg, 2002; ch. 3-5]. Finally, we offer some concluding remarks.

Emdros: Nature and scope

Emdros is a database management system for managing a particular kind of data, namely “annotated text”. Annotated text can be described as “text plus information about that text.” Text is fundamentally different in nature than, say, employee records or databases of movie information, due to some properties of text which are not usually present in data which fits easily into the relational database model. Three such properties of text include: First, the linear nature of text, in that text is meant to be read and understood in sequences of varying length. Second, the embedded nature of text, in that language can be seen by linguists, writers, publishers, and others to exhibit hierarchical structures, including the “word --> clause --> phrase --> sentence” linguistic hierarchy, and the “sentence --> paragraph --> section --> chapter --> book” compositional hierarchy familiar to readers and writers. Third, the phenomenon of “gapping”, or “distant relations”, whereby one part of a sentence – such as the preceding – is separated from another by a parenthetical clause or phrase (as just exemplified). All three phenomena – sequence, embedding, and gapping – are properties of


text which call for special handling in a database, and Emdros caters to each of these properties, as well as other properties of annotated text, in meaningful ways. It does so by implementing a database model and a query language which make it easy to think about, implement, query, and use the textual data.

Thus Emdros implements a particular database model of text, which makes it easy to think about text in database terms. The database model is called the “EMdF model”, and is a descendant of the MdF model described by Crist-Jan Doedens in his PhD dissertation [Doedens, 1994, chs. 2-4]. Similarly, Emdros implements a query language called “MQL”, which again makes it easy to query annotated text in meaningful ways. The query results of MQL fit naturally with the EMdF model, just as the query results of an SQL query fit naturally within the relational model [Codd, 1980]. The MQL language is a reformulation and implementation of the QL language also described by [Doedens, 1994, ch. 6].

Emdros also implements other layers, including a “Full Text Search engine”, built on top of MQL. The “Full Text Search” engine provides proximity searches, exact phrase search, optional trunkation of search terms, and other features.

Another layer provided by Emdros is a “harvest layer”, which makes it easy to “play back” parts of documents stored in an Emdros database. For example, an XML, HTML, PDF, or MS Word document, having been stored in an Emdros database, can be easily “rendered” back as either HTML, XML, JSON, RTF, LaTeX, or whatever notation one wishes to employ. The format to be output is described in a “style sheet”, which tells the harvest layer what to do with textual objects in the database such as “paragraph”, “sentence”, “token”, or whatever object types the database designer has chosen to implement in the database application under construction.

A relational database management system does not dictate what tables to put in a particular database. It merely provides a database model (the relational model) within whose parameters (tables, columns, column types, primary keys, foreign keys, etc.) the data must be modelled in a schema. The schema, in turn, provides the parameters within which the actual data can be used to populate the database. A relational database management system also usually provides a query language (most likely SQL) in which to define the schema, create the data, update the data, query the data for retrieval, and delete the data.

Similarly, Emdros does not dictate which kinds of textual annotations of textual elements to store in a database. Instead, it provides parameters within which those textual elements and textual annotations can be defined in a schema, and entered into the database as data. Similarly, Emdros also provides a query language (the MQL query language) for defining the schema, creating the data, updating the data, querying it in meaningful ways for retrieval, and deleting the data. Thus the user is free to choose, for example, whether to include a “sentence” object type, and whether to store the text as individual graphemes, morphemes, words, sentences, or even higher-level units as the basic unit of text. The construction of the schema, with its concomitant choices for the determination of the nature of the textual data, is up to the programmer and/or database designer who uses Emdros. For an overview of these aspects of Emdros, please see [Sandborg-Petersen, 2008], as well as the Emdros documentation available from the Emdros website, http://emdros.org/.

Emdros is not meant to be used in isolation. It implements some layers which has the potential greatly to reduce the code-base of any system having to store text in a database. Thus Emdros plays its part in a larger system, for example an application sitting on the user's desktop, on an iPad or iPhone, on an Android device, or even on a server. The two great advantages of Emdros are: First, its ease of use for the programmer of the system dealing with text, whose task is made much simpler and cleaner. And second, its speed. Both of these


aspects have been described in [Petersen, 2005, 2006a, 2006b] and [Sandborg-Petersen, 2008]. For a concise introduction to Emdros, see [Petersen, 2004].

Emdros is Open Source under the GNU Public License version 2, and is available for free download from the Emdros website.

Case studies

In the following, we describe the nature and architecture of each of the three systems under case-study. We first describe the PLOTLearner, then the Kaj Munk Study Edition, and finally the Kaj Munk Virtual Research Laboratory.

PLOTLearnerPLOTLearner is a tool developed within the EuroPLOT project which aims to make it easy to create, reuse, and consume Persuasive Learning Objects within the domain of language learning. It builds on a number of ideas, including the following: First, that language learning can be driven by particular texts, where those texts have been analysed linguistically. Second, the idea that the quiz format, with its questions and required answers, facilitates learning. An early version of PLOTLearner has been described in [Tøndering, 2009].

The idea is to give the learner a piece of a text, along with a number of questions which the learner must answer about the text. The questions are taken directly from a database of linguistically analysed or annotated text, embodied in an Emdros database. It is the task of the learner to answer the questions correctly, and thus gain greater insight into the system of the language being studied. For example, questions can relate to the parts of speech, tense, aspect, number, gender, meaning, and other lexical and morphological aspects of individual words. Or questions can relate to features of phrases or clauses, depending on the underlying database.

Figure 1. Architecture of PLOTLearner

PLOTLearner has been designed to be language-independent and text-database driven. Thus a number of databases for various languages have been published alongside PLOTLearner, including Ancient Hebrew, Ancient Greek, and modern German. There is a “facilitator's view” (or “teacher's view”), which deals with designing questions and saving them in a separate file as a persuasive learning object. The other view deals with “playing back” the questions and


asking for answers in a “student's view”. The “facilitator's view” depends on the underlying Emdros database for drawing information for the questions from the annotated text.

Thus Emdros it situated at the bottom of the architecture of the application, as depicted in Figure 1. The architecture is, as can be seen, a fairly standard “Model – View – Controller” architecture, with a database layer underneath in the form of Emdros.

Kaj Munk Study EditionThe Kaj Munk Study Edition in an earlier form, its content, its rationale, and its architecture are described in greater detail in [Sandborg-Petersen, 2008]. Here, the following description will suffice for our purposes. The Kaj Munk Study Edition has been further developed within the EuroPLOT project. It is in essence a “Kaj Munk Browser”, which builds on the rather self-evident idea that in order to get to know an authorship, one has to have access to the texts constituting the authorship. Thus the Kaj Munk Study Edition software contains and displays the works of Kaj Munk, an important cultural figure in Denmark and even in Europe in the period between World War I and World War II. The software provides easy access to not only the works of Kaj Munk, but also a wealth of explanatory notes, introductory material, pictures, and other information about Kaj Munk and his works. Both simple and advanced search facilities are included, there are “backwards” and “forwards” buttons, and in general, the user experience is much like that of a “web browser”, only it is the works of Kaj Munk that are “browsed”, rather than the web.

The wealth of information provided by the Kaj Munk Study Edition is shown to the user in a number of panes driven by custom HTML widgets. The left part of the screen displays the text being read. The right part of the screen can be chosen by the user display the list of works, ancillary information, explanatory notes, or search results.

Figure 2. The architecture of the Kaj Munk Study Edition software

As can be seen in Figure 2, the architecture of the Kaj Munk Study Edition software also has Emdros as its underpinning. The code base of the Kaj Munk Study Edition is very small considering the many features. In fact, it only comprises about 10,200 lines of C++, not including the aforementioned HTML widget. Of these 10,200 lines, most are taken up by the View and the non-database-facing parts of the Controller. Only about 1,600 lines are


necessary for retrieving and assembling all of the various kinds of displayable information from the databases of diverse kinds. Thus Emdros is seen to be easy to use for the programmer, and to assist in reducing the number of necessary lines of code to a large extent, as compared to what would otherwise have been necessary.

Figure 2 also shows the inner architecture of Emdros itself: At the bottom, there is a storage backend; in this case, it is SQLite 3, but other backends are supported such as PostgreSQL, MySQL, SQLite 2, and the author's own proprietary “Bit Packed Tables” or “BPT” engine. The latter makes deployment of Emdros feasible and useful on low-powered devices such as iPods, iPads, iPhones, and Android devices, due to the speed of the BPT engine.

The next layer of Emdros is the database model layer, embodied in the EMdF layer. This in turn is called upon by the query language layer, the MQL layer. The application is able to use MQL directly, or use any of the other layers, including the harvest layer and the Full Text Search layer.

The Kaj Munk Study Edition implements its own Full Text Search layer, for historical reasons, but could easily be adapted to use the Full Text Search layer of Emdros, thus reducing its code-base even further. The Controller layer of the Kaj Munk Study Edition is a thin layer between the upper View (GUI + HTML widget) layer, while the Text Retrieval layer is a very thin layer on top of the Harvest layer of Emdros.

Thus the architecture of the Kaj Munk Study Edition software is similar to that of the PLOTLearner: A Model – View – Controller application with Emdros serving as a backend. As has been pointed out, the code base of the Kaj Munk Study Edition is rather small compared to the wealth of features it provides, and this is due to the power and ease of use of the undergirding database engine.

Kaj Munk Virtual Research Laboratory

Figure 3. The architecture of the Kaj Munk Virtual Research Laboratory

The Kaj Munk Virtual Research Laboratory is a web-based collaborative research application, which aims to gather researchers around the texts of Kaj Munk. It provides the literary, historical, or theological researcher with tools to author explanatory notes to the works of Kaj


Munk, and to anchor these notes in the text. The Laboratory also provides the researcher with social tools such as the ability to comment on the notes of other researchers, thus keeping the scholarly debate and dialogue open and transparent. The notes themselves can also be viewed in various ways, both situated within the texts and aggregated over many texts at once, thus providing a better overview of the corpus of annotations having been written. Finally, the Laboratory provides access to the works of Kaj Munk, both in terms of displaying the works, and in terms of searching them.

The web application itself is a very thin layer built on top of Emdros, using the same databases that are used in the Kaj Munk Study Edition. The architecture can be depicted as in Figure 3.

In a sense, the Kaj Munk Virtual Research Laboratory possesses the simplest architecture of the three systems under case-study, in that it does not have a model in and of itself – Emdros is enough to provide the model entirely. Thus it implements a View and a Controller, and lets Emdros play the role of both the database backend and the model.

Persuasive nature of the case-studies

As described by B.J. Fogg in Chapters 3 through 5 in [Fogg, 2002], Persuasive Design encompasses three ways in which a computer system can be persuasive, as embodied in Fogg's “Functional Triad”, also described in [Fogg, 1998]:

1. First, by being a “Persusaive Tool”, which “does” something for the user.

2. Second, by providing “Persuasive Media”, in which the user is encouraged to consume information, for example.

3. Third, by being a “Persuasive Social Actor”, in which the computer either facilitates and mediates social interaction between humans, or interacts socially with humans by itself.

Fogg's Functional Triad can be depicted as in Figure 4.

Figure 4. The Functional Triad

As summarized in Table 1, the three systems under case-study exhibit some or all of these persuasive characteristics in Fogg's “Functional Triad”.

Table 1. The Relationship between the Cases and the Three Legs of Fogg's Functional Triad

System Tool Media Social Actor

PLOTLearner X XKaj Munk Study Edition XKaj Munk Virtual Research Laboratory X X X


The PLOTLearner system can be characterized by both being a “Persuasive Tool” – because it persuades the user to do something in the task of learning a language – and by providing “Persuasive Media”, in that PLOTLearner provides the opportunity to consume both some text from the language being studied, and an analysis of these texts. Notice that Emdros plays a central role in the implementation of both the “persuasive tool” aspect, and the “persuasive media” aspect, given that both the quizzes and the texts from which they are derived have their origin in Emdros within the system.

The Kaj Munk Study Edition provides “Persuasive Media”-style access to the texts of Kaj Munk by making it very easy to study the works of his authorship. The program is itself not a social actor, nor does it facilitate social interaction between students of the works of Kaj Munk. The program does not really function as a persuasive tool either, since it does nothing particularly persuasive or special to assist the user in his or her study of the texts. That is, there is no set agenda through which the order of the steps being taken by the user in order to study the works of Kaj Munk can be carried out. Thus the concept of persuasive “tunnelling” does not apply, for example. Notice again that Emdros has a central role to play in the implementation of the persuasive media, in that almost all of the various kinds of media being offered for consumption by system have their origin in Emdros.

The Kaj Munk Virtual Research Laboratory, on the other hand, exhibits all three legs of Fogg's Functional Triad. First, it is a tool which persuasively helps the user perform a task, namely the annotation of the works of Kaj Munk. Second, it provides access to “persuasive media”, by making it easy for the user to consume both the works of Kaj Munk and the annotations connected to the text. Third, it is a “persuasive social actor”, in that it facilitates easy social interaction between the various researchers who use the system, as well as being a social actor in its own right in its persuasive messages to the user, which sometimes are of an encouraging or even forbidding nature. Notice yet again that Emdros, being the model component of the system, plays a very central role in all three persuasive aspects of the system.

Not all persuasive learning systems will be able to benefit from Emdros. Systems which do not deal with text as the primary content in their persuasive learning objects may not be able to benefit much from Emdros, being, as it is, primarily a text database engine. However, as can be seen, Emdros is able to facilitate the implementation of all three legs of the Functional Triad inherent in Persuasive Design, probably reducing the necessary size of the code base. Thus persuasive learning systems of many kinds will be able to benefit from using a solidly theoretically founded text database engine such as Emdros. In addition, any system whose primary functions include giving access to, searching, creating, or otherwise dealing with text plus information about that text in a database are likely candidates for systems that could benefit from incorporating Emdros.

Conclusion

In this paper, we have provided an overview of the Emdros text database engine for annotated text, and have developed three case-studies of persuasive learning systems which incoporate Emdros as an architectural underpinning. The three systems have been described as regards purpose, nature, functionality, architecture, and as regards the three legs of the Functional Triad og Persuasive Design described by [Fogg, 2002]. We have argued that Emdros is capable of undergirding all three aspects of any persuasive learning system whose primary content includes text plus information about that text. We have also argued that the power and ease of use of Emdros makes it likely that the size of the code base necessary to accomplish the goals of such systems can be reduced by incorporating Emdros, as compared to systems


which use other forms of databases. This is due to the fact that Emdros caters specifically to some aspects of text which are hard to implement in traditional database management systems, including the phenomena of sequence, embedding, and gapping.

Acknowledgements

The author wishes to thank Claus Tøndering for important insight into the PLOTLearner architecture, provided through personal communication.

References

Codd, E.F. (1980). Data models in database management. In International Conference on Management of Data, Proceedings of the 1980 workshop on data abstraction, databases and conceptual models, Pingree Park, Colorado, USA, pages 112–114, 1980. ISBN 0-89791-031-1.

Doedens, Crist-Jan (1994). Text databases: One database model and several query languages. PhD dissertation, Universiteit Utrecht, Netherlands.

Fogg, B.J. (1998). Persuasive computers: perspectives and research directions. In CHI '98. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Los Angeles, California, USA, pp. 225-232. ACM Press / Addison-Wesley Publishing Co., New York, NY, USA.

Fogg, B. J. (2002). Persuasive Technology: Using Computers to Change What We Think and Do. Morgan Kaufmann.

Petersen, Ulrik. (2004).Emdros – a text database engine for analyzed or annotated text, In ICCL, Proceedings of COLING 2004, held August 23-27 in Geneva. International Committee on Computational Linguistics, pp. 1190—1193.

Petersen, Ulrik. (2005). Evaluating corpus query systems on functionality and speed: TIGERSearch and Emdros. In: Angelova, G., Bontcheva, K., Mitkov, R., Nicolov, N. and Nikolov, N. (eds): International Conference Recent Advances in Natural Language Processing 2005, Proceedings, Borovets, Bulgaria, 21-23 September 2005, pp. 387—391.

Petersen, Ulrik. (2006a). Querying both Parallel and Treebank Corpora: Evaluation of a Corpus Query System. In: Proceedings of LREC 2006.

Petersen, Ulrik. (2006b). Principles, Implementation Strategies, and Evaluation of a Corpus Query System. In: Yli-Jyrä, Anssi; Karttunen, Lauri; Karhumäki, Juhani (eds), Finite-State Methods and Natural Language Processing 5th International Workshop, FSMNLP 2005, Helsinki, Finland, September 1-2, 2005, Revised Papers, Lecture Notes in Computer Science, Volume 4002/2006, Springer Verlag, Heidelberg, New York, pp. 215—226. DOI: 10.1007/11780885_21.

Sandborg-Petersen, Ulrik (2008). Annotated text databases in the context of the Kaj Munk archive: One database model, one query language, and several applications. PhD dissertation, Department of Communication and Psychology, Aalborg University, Denmark. Available from http://ulrikp.org/PhD/

Tøndering, Claus (2009), 3ET – An Automatic Tool for Grammar Training, Hiphil, Volume 6 [http://www.hiphil.org] Accessed 30.06.2013.


Application Domains

Persuasive Design can be applied in a variety of applications, and different technologies can appear

most suitable for specific application needs. Two examples will be presented here: Gram-Hansen has

used mobile learning technologies for teaching about Danish literature. Bertel and Rasmussen have

employed entertainment robotics in Autism therapy and education.


Motivating the interest in Danish literature with Mobile Persuasive Learning

Sandra Burri Gram-Hansen, Karina Dyrby Kristensen Department of Communication and Psychology, Aalborg University, Denmark

[email protected] , [email protected]

Lasse Burri Gram-Hansen Royal School of Library and Information Science, University of Copenhagen, Denmark

[email protected]

This paper analyses and discusses the potential of Mobile Persuasive Learning (MPL) in

relation to learning scenarios that involve complex and interdisciplinary learning material.

A specific example of MPL is presented, which has been developed with the intent to

motivate the interest of the life and works of Danish author and playwright Kaj Munk. A

Persuasive Learning Design (PLD) is tried in a specific learning scenario that aims to

introduce the history of Kaj Munk to students in lower secondary education in Vester

Hassing in Northern Jutland. The methodological background for the chosen scenario is

described and evaluation activities are presented and discussed and it is argued that while

the topic is possibly too complex for the particular age group of students chosen,

evaluation feedback and researcher observations point towards a significant potential in

further developing MPL-designs in a school context.

Keywords: Mobile Persuasion, Collaborative Learning, Kaj Munk, e-PLOT, Mobile

Persuasive Learning

Introduction

This paper presents and discusses a test project concerning MPL in a small Danish village,

where a group of students were involved in a number of learning acitivities regarding the

historically significant Danish vicar Kaj Munk.

Kaj Munk served as vicar in Vedersø in Western Jutland, and was assassinated by Gestapo in

January 1944 after having aired his resistance against the German occupants through his

sermons and his literary works. As a result, Kaj Munk has become an important figure not

only in Danish literature, but also in Danish history, and the mediation of his life and works

receives attention both in learning and in mediation of cultural heritage. One of the interesting

aspects of the works of Kaj Munk is the fact that many of his dramas and other texts are

closely related to specific places – mainly in Denmark and very often in the area around

Vedersø in western Jutland where he lived and served as vicar. The close connection between

many of Kaj Munk’s texts and specific locations in different parts of Denmark, constitutes the

primary reason why MPL may show particular potential in this work case (Gram-Hansen &

Gram-Hansen, 2013) (Gram-Hansen, 2013b).

The Kaj Munk example springs from research conducted within the EU funded e-PLOT

project. E-PLOT aims to explore the theoretical and practical overlap between persuasive

design and learning technologies, through further development of two existing technologies;


PLOTLearner and PLOTMaker. It has been argued that mobile devices hold potential which

may ease the task of incorporating the surroundings in a design (Fogg & Eckles, 2007)

(Bødker, 2012), and as a result PLOTmaker now provides teachers with the ability to create

not only traditional but also mobile learning objects which can be executed through

smartphones and tablet computers. In order to test both types of learning objects, the

Persuasive Learning Design (PLD) presented in this paper, includes both traditional and

mobile learning objects all of which are linked to an overall intended outcome.

The trial took place at the local school in a small village in Northern Jutland called Vester

Hassing. The link between location and learning material in this is case exemplified by a

historical anecdote describing how Kaj Munk once agreed to give a sermon in Vester Hassing

church, but only after having received written notice from the village snow bailiff, letting him

know that the local pastor was in fact a very funny man - a notice which Kaj Munk found

funny enough itself to make him agree to visit the village. The anecdote itself is not of major

significance to Kaj Munk’s works in general, but it does provide an example of the mentality

of Kaj Munk which might facilitate the student’s understanding of him as a historical figure.

Furthermore, the fact that the home village of the students was somehow related to Kaj Munk

was also expected to motivate the students’ interest in the topic - especially as Vester Hassing

is a small village with little significance in the overall history of Denmark. The persuasive

learning design in Vester Hassing was divided into three phases, followed by an evaluation

session involving all participating teachers, observers and the entire group of students. The

inclusion of both traditional and mobile learning objects, sought to constitute a multimodal

learning experience with a strong emphasis on collaborative learning.

From Geocaching to Mobile Persuasive Learning

A 2006 pilot study from Thorning elementary school, Denmark, used mobile technology to

enable teaching outside the classroom (Thomsen, 2008). Using the location-based GPS-hunt

Geocaching, teachers motivated students to learn different subjects in appropriate

surroundings leveraging the students’ fascination with technology to spark their interest.

Generally, the students were successfully motivated to engage in learning via location-based

mobile technology, and even in situations when the educational benefit was minimal, the

students were still motivated by the idea of doing classes outside the classroom. However, the

teachers learned that they needed to carefully prepare material for this specific type of

teaching. When the content was planned more thoroughly, the geocaching activity could be

used to motivate students of all ages and in a wide variety of subjects.

Subsequent research has argued that the dynamics of Geocaching hold an interesting

persuasive design potential, as Geocaching combines elements such as location and context

aware technologies with elements of competition, rewards and a number of persuasive design

principles (Gram-Hansen, 2009). Ongoing research in digital dissemination of cultural

heritage suggests that the dynamics of Geocaching holds the potential to motivate an interest

in areas of cultural heritage, which may not receive much attention otherwise. The research is

exemplified by a number of Geocaches located in the area around Vedersø where Kaj Munk

served as vicar.

The results of both the pilot study and the ongoing investigation of Geocaching in a PD

context in Vedersø, serves as a primary source of inspiration for the Persuasive Learning

Design (PLD) in Vester Hassing. The overall plan for the design was to create a multimodal

learning experience where an appropriate balance was established between the learning

material, the intended learning outcome and the learning context. Of particular importance

was the understanding that being at a specific location when receiving learning material,


influences the way students perceive the material. In consideration of the time that has passed

since WWII, the students have no natural understanding of the age in which Kaj Munk’s

literary works were produced. As a result, the learning design sought to not only present the

learning material in an appropriate and motivating manner, but also to enable the students to

create their own references to locations and events which were important in that time.

Mobile Persuasive Learning in Practice

Previous research has argued that the unique claim of persuasive design is not tied to certain

technologies, but rather depends on how these technologies are applied within a specific

context. As such, persuasive design becomes the discipline that creates the link between

technology and context, based on a strong focus on appropriateness within that context.

Rather than constituting an approach to technology design which offers little novelty when

related to more established research fields, persuasive design becomes a Meta layer, where the

context is explored in a wider sense, and where the focus of the designer is to establish a link

between the context, the persuasive intention and available technology while remaining a

recurring ethical perspective (Gram-Hansen, 2013a; Gram-Hansen, 2013b). When related to

learning design, this new and more nuanced understanding of persuasive design

acknowledges the various recognised approaches to learning, which are already applied in a

variety of setting. The PD approach supports these approaches by adding an overall reflective

perspective, which motivates teachers to consider ways to actively adapt the surroundings in

the learning design.

One of the challenges for teachers who work with a constructivist approach to learning, is

argued to be contradiction between intentions related to the learning scenario and the physical

frames of the setting (e.g. an auditorium does not motivate student activities; see (Leijon,

2012). By its wide and context-oriented perspective, persuasive design may potentially

provide teachers and learners with new approaches to teaching and learning, which may help

overcome the challenges related to physical surroundings. As a result, the persuasive learning

design describes the overall understanding of the intended learning scenario, and the learning

technologies applied within the learning design become facilitators for reaching the intended

outcome. Thereby it is still agreed that a technology is only persuasive, when applied

appropriately within the intended use context.

As described above, the described understanding of persuasive learning designs has been tried

in a specific learning scenario that aimed to introduce students in lower secondary education

in Vester Hassing in Northern Jutland, to Kaj Munk. The 18 students that participated in the

trial were 13-14 years of age and did not have any previous knowledge of Kaj Munk. From

films and literature, they did however know a little about WWII, and did as such have some

understanding of the dynamics that influenced the world when Kaj Munk was murdered.

PLOTMaker enables teachers to produce both traditional and mobile learning objects. In order

to test both types of learning objects, the persuasive learning design developed for the

students in Vester Hassing, included both traditional and mobile learning objects all of which

were linked to an overall intended learning outcome. The use of mobile learning objects were

of particular relevance to the Kaj Munk case due to the connection between historical events,

literature and specific locations, which is a particular distinction for this work case.

The overall persuasive learning design was widely inspired by the previously mentioned

results from applying geocaching in digital mediation of cultural heritage. In particular, the

notion of Mystery Caches, where the users must solve a puzzle in order to find a specific

location, was a motivating factor in the development of the learning design.


The overall approach to the development of the persuasive learning design is based on the 6

step guide to persuasive learning which has been developed in relation to the e-PLOT project

(Kristensen, 2013). The 6 step guide combines an acknowledged Danish approach to learning

designs with Fogg and Eckles’ 8 step approach to persuasive design, and seeks to provide

teachers with a step-by-step framework for developing persuasive learning designs. The guide

approaches learning design preparation through two different phases, the first one placing a

focus on the learning scenario, and the second phase focusing more specifically on the

learning material. Both phases include contextual reflections based on the rhetorical notion of

Kairos, and both phases undergo a recurring evaluation process, in order to ensure that the

final learning design meets the requirement of an appropriate balance between intended

outcome, technology and context.

Put to practice, the guide served as a supporting framework throughout the design process,

ensuring that particular attention was drawn towards the context, and more specifically the

balance between student’s ability, the technologies used, and the intended learning outcomes.

With reference to the wider understanding of the concept Kairos and the notion of context

adaptation (Gram-Hansen, 2013a), the influence of persuasive design in this particular

learning design is defined by the reflections regarding how the student’s normal surroundings

and media can be applied in the learning scenario. As the intended outcome of the learning

design is for the students to learn about Kaj Munk, it is important that the technologies do not

outshine the learning material by drawing all attention. Nor must the technology appear too

complex, as the students may then focus more on how to use the devices and less on what the

technologies are mediating. Instead, the technologies and the learning objects mediated

through them must function as facilitators that underpin important arguments at appropriate

moments, for instance when the students are at a specific location, or as a way to follow up on

a more general introduction. These reflections proved themselves to be particularly important

due to the age group of the students. Whilst older students may be able to distance themselves

from the excitement of new technologies in a learning scenario, the students in Vester Hassing

appeared be more easily distracted by new devices.

Prior to the trial, the students were told about the e-PLOT project and explained that their

input and comments about both the learning experience as a whole and the different

technologies is both welcome and important to the researchers in the project. They were also

informed about the overall plan for the day; that they would be doing a combination of group

work in the classroom, and a trip around their town, during which a tablet computer would

guide them and supply them with additional learning material. They were informed that an

observer would accompany them, both in the classroom and en route, but that they were

expected to work independently. It was heavily emphasised that from a research perspective,

the aim of the day was to explore how the students respond to this approach to learning, and

to learn more about what motives the students.

In practice, the learning design consisted of three different phases followed by an evaluation:

1. The introduction phase, during which the students were given a general introduction

to Kaj Munk and the intended learning outcomes of the day. The theoretical

framework in e-PLOT includes reflections regarding the notion of Constructive

Alignment in relation to outcome based learning, which emphasises the importance of

letting the intended learning outcome be known to both teachers and students (Biggs

& Tang, 2007; Gram-Hansen, 2012). The intended outcomes for the students were as

follows:

The students will gain a basic understanding of Kaj Munk as a historical figure and

of the role he played during WWII.


The students will be able to identify the link between Kaj Munk’s works and

specific locations in Denmark

This phase took part as a traditional classroom setting, where the material was

presented through PowerPoint. The introduction was kept very short (approx. 15

mins.), and was mostly used for practical information. The introduction to Kaj Munk

was limited to three points:

Kaj Munk was a Danish Vicar

He was a poet and a playwright

Gestapo murdered him during WWII.

As one of the aims of the trial was to evaluate the potential of persuasive learning

objects, it was considered important to let the students know as little as possible before

engaging in the learning material.

2. The classroom phase, during which the students worked in smaller groups (4-5

students in each group), and engaged in more specific exercises. All groups were

presented with a worksheet with different questions about Kaj Munk and WWII, and

in order to answer these questions the students had to explore a number of Kaj Munk

related Learning Objects. Originally, the intention behind the exercise was that by

answering the questions correctly the students would be provided with a series of GPS

coordinates that would determine their route around Vester Hassing. However, in

order to ensure that not all groups would be on the route at the same time, it was

decided to simply dispatch groups every 10 minutes, and let the groups which returned

back to the classroom first, complete the worksheet upon their return.

3. The Mobile Learning phase, which took place on a tour through Vester Hassing. As

the groups were dispatched, they were given a 7” tablet that had been prepared to

execute the learning material. The students were sent out on a specific route to explore

the village and amongst other things discover how buildings that no longer serve a

specific purpose, carried historical significance. For instance, the students were sent

by the old train station in Vester Hassing, and asked to solve a photo challenge, whilst

learning about the time when members of the Danish resistance in WWII blew up the

railroad in Vester Hassing in an act of sabotage. While at the specific location, the

students were presented with a photo of the station as it appeared during WWII, and

asked to use their cameras in their mobile phones to capture a new picture of the

station, in the same angle as the old picture.

The learning material and the tasks presented to the students during the tour of the

town, was presented through mobiles GLOs, which were activated by scanning a QR

code. The decision to activate the Mobile GLOs with QR codes rather than GPS, was

based on the students working in groups rather than individually. The GPS option

would enable the learning material to be activated on specific locations, however the

QR codes held the benefit of engaging all members of the group in searching for the

location, rather than just the person holding the tablet.

Evaluation and reflection

Once having completed the 3rd phase, the students reconvened in the classroom for a closing

evaluation session. Through a combination of quantitative and qualitative evaluation methods,

the aim was to evaluate not only the learning potential of the design, but also the


persuasiveness of the experience as a whole. Evaluation the persuasive potential of the design

called for careful consideration, as the students due to their age group were believed likely to

respond positively to the learning experience, simply as a result of the activities being

something new and technology-based.

In practice, the evaluation phase consisted of a number of activities which sought to be

inclusive towards all students, and which allowed students to respond anonymously to

questions regarding the learning experience.

In order to assess if the students had met the intended learning outcomes, the evaluation phase

started with a group competition in terms of a quiz. A number of questions were read out one

at a time, and for each question, the groups were given a moment to quietly discuss their

answer and write it down on the quiz form. The questions were all based on information the

students had been presented with during the day, either during the classroom phase or on the

route in Vester Hassing. The groups were informed that they were not allowed to use their

notes from the earlier phases to answer the questions; they were however allowed to use the

pictures they had taken while solving the tasks of the MLD. If the photo challenges had been

solved correctly, their pictures would include some of the answers to the quiz.

Although the students had only been told very little about Kaj Munk in the beginning of the

day, all groups performed impressively well in the quiz. The quiz included questions such as

“Why was Kaj Munk murdered”, “Why did he agree to speak in Vester Hassing”, “How many

of Kaj Munk’s literary works can you mention?” and “What is Kaj Munk trying to say in his

song about The Blue Anemone?”, and did as such require the students to not only know facts

about Kaj Munk, but also be able to reflect upon the material they had worked with. Results

showed that even though the students had not been told much about Kaj Munk before

engaging in the different phases, three out of five groups were able to answer all ten questions

correctly, and the remaining two groups each had one wrong answer.

The positive results of the quiz may be credited to the students’ engagement in the MPL tour

around town. One of the particularly interesting observations made in several of the groups,

were that when reaching one of the intended location, one group member would read aloud

the content of the learning object, and the remaining group members would appear fairly

uninterested. The technology itself did not motivate much interest, and in some cases no one

really wanted to be the one in charge of the device. In spite of showing such little enthusiasm

on location, observers noted that as the groups were walking from one location to the next,

they would discuss and reflect upon the material they had just received, and in some degree

translate it into something they could all understand. As such the application of mobile

learning objects appeared to motivate an important degree of collaborative learning, which all

members of the group were able to engage in and benefit from.

As mentioned, the evaluation of the persuasiveness of the PLD had called for careful

methodological considerations, as the students were perceived to potentially be more easily

influenced by the situation. The evaluation was approached form a qualitative perspective, as

it was considered of greater importance to gain a deeper understanding of how the students

experienced the PLD, and whether this approach to learning had any motivating effect. The

overall approach to the qualitative evaluation is based on phenomenography (Ashworth,

2010), and sought to influence the students as little as possible, and to allow them the

anonymity to respond critically if necessary.

Much was gained from the observation studies, and in particular the walk along and in situ

interviews during the Mobile Learning Phase provided valuable insight to the potential of the

learning design and to the actual application of mobile devices in a learning scenarios. For

instance, the mentioned lack of enthusiasm from the students as they were on the route was


surprising, as expectations were that the use of tablet computers in itself would be a

motivating factor, as this is not yet widely applied in Danish schools. The original concern

that the students might be so engaged in applying mobile devices for learning, that the device

would outshine the learning material was proven to be unnecessary. Contrarily it may be that

as the students are born as digital natives and with mobile devices as a natural element in their

everyday surroundings, the future of MPL calls for much richer ways to present learning

material. Technologies themselves do not impress the future users, so the design of the device

applications must constitute an engaging and intriguing experience. Particularly if the design

is to hold persuasive potential, as likeability is argued to be an important motivational factor

(Fogg, 2003).

In order to supplement the observations made during the day with specific input from the

students, the day was finished with a couple of evaluation activities that enabled the students

to submit individual feedback. All students were asked to fill out a questionnaire regarding

their attitude towards both Kaj Munk and the subject of history in general, and finally all

students were given a blue and a yellow piece of paper. On the blue piece of paper they were

asked to complete the sentence “I learned a lot today because…” and on the yellow piece of

paper they were asked to write down three adjectives, which described their impression of the

day. The dominant adjectives entered on the yellow piece of paper were “Fun”, “Exciting”

and “Different”. A majority of the students credited the MPL experience, the collaborative

learning and the high level of physical activity as the main reasons why their learning

outcome had been good. On student did comment that he/she was unsure about the actual

learning outcome, as many of the details about Kaj Munk would most likely soon be

forgotten.

Results of the questionnaire showed that a majority of the students found that the PLD

experience motivated them to learn more about history in general, however only 4 out of 18

students replied that the experience motivated them to learn more about Kaj Munk. 4 students

replied that they did not feel motivated to learn more about Kaj Munk, and 10 students replied

that the experience did not influence their interest in Kaj Munk in either direction. These

results were not as positive as hoped, but also not entirely unexpected. Although a majority of

the students responded positively with regards to being motivated towards history lessons in

general, one of the aims of the learning design was to motivate an interest in the life and

works of Kaj Munk and this was not entirely achieved by the learning design. One reason for

this may be that although the students do learn about WWII, the history about Kaj Munk and

in particular his literary works constitutes material that may be too complex for students of

this age to comprehend. Nonetheless, the benefit of testing the PLD with Kaj Munk material

remains that as the students have not previously been introduced to Kaj Munk specifically,

they students commonly engage in the experience with no knowledge about the material they

are to work with, This taken into consideration, it can be argued that the learning potential of

the design is considerable.

In summary, there is no empirical support for claiming that the MPL trial in Vester Hassing

directly motivated the students’ interest in the subject of Kaj Munk. However, the students

showed that they were able to learn about a complex subject through the combination of

learning objects and different learning activities. Furthermore, the trial provided a valuable

experience in the use of mobile persuasive learning objects in practice. Based on the

experiences of this trial, it would be interesting to explore the outcome of a similar PLD, in

which the learning material is better matched to the specific user group.


References

Ashworth, P. a. L., Ursula. (2010). Achieving Empathy and Engagement: A practical approach to the design, conduct and reporting of phenomenographic research Studies in Higher Education (pp. 295-308): Routledge.

Biggs, J., & Tang, C. (2007). Teaching for Quality Learning at University (3rd ed.): Society for Research into Higher Education & Open University Press.

Bødker, S. a. E. C. (2012). Poetry in motion – Appropriation of the world of Apps. The 30th European Conference on Cognitive Ergonomics, ECCE '12.

Fogg, B., & Eckles, D. (2007). Mobile Persuasion - 20 Perspectives of the Future of Behavior Change: Persuasive Technology Lab, Stanford University.

Fogg, B. J. (2003). Persuasive Technology - Using Computers to Change What We Think and Do. San Francisco: Morgan Kaufmann Publishers.

Gram-Hansen, L. B. (2009). Geocaching in a persuasive perspective. The 4th Conference on Persuasive Technology, Claremont, CA.

Gram-Hansen, S. B. (2012). Deliverables Report D.3.3, Persuasive Learning Designs: Aalborg university.

Gram-Hansen, S. B. (2013a). Persuasive Design - A Matter of Context Adaptation. IWEPLET 2013 - Workshop on Persuasive Technology in Learning and Teaching, Paphos, Cyprus.

Gram-Hansen, S. B., & Gram-Hansen, L. B. (2013). On the role of Ethics in Persuasive Design. ETHICOMP2013, Kolding.

Gram-Hansen, S. B. a. L. B. G.-H. (2013b). On the role of ethics in Persuasive Design. Ethicomp 2013, Kolding.

Kristensen, K. D. (2013). A 6 Step Guide to Persuasive Learning ECTEL, Paphos, Cypres. Leijon, M. (2012). The room in higher education – a space for learning? DfL2012 – The third Designs

for Learning conference, Copenhagen. Thomsen, T. B. (2008). GPS i Undervisningen. Thorning Skole: Thorning Skole.


PEERs at Play: A Case Study on Persuasive Educational and Entertainment Robotics in Autism Education

Lykke Brogaard Bertel, Dorte Malig Rasmussen Danish Technological Institute, Odense, Forskerparken 10F 5230 Odense M

[email protected] , [email protected]

This paper introduces the notion of Persuasive Educational and Entertainment Robotics

(PEERs) as the intersection between Persuasive Design, Human-Robot Interaction and

Didactics and investigates the persuasive potential of PEERs in special needs education

through an empirical study on the implementation and use of the robot seal PARO at a

school for children with autism. Results include a categorization of user-developed

concepts of PEER-supported interventions in special needs education as well as

persuasive design principles that emerges from the comparison and combination of design

strategies within Persuasive Technology, Human-Robot Interaction and Didactics.

Keywords: Persuasive Design, Human-Robot Interaction, Autism Education

Introduction

Robots are increasingly implemented as teaching aids and assistants at universities, secondary

and even primary schools to increase interest STEM teaching (Mataríc, Koenig & Feil-Seifer,

2007 and Majgaard & Misfeldt, 2010) and language learning (Han & Dongho, 2009) by

providing a platform for experimentation, reflection and collaboration. The application of

robots to motivate social interaction and communication in autism education has also gained

increased interest in recent years (Dautenhahn et al, 2009, Lee, Breazeal & Picard, 2008,

Robins et al, 2009, Kozima, Nakagawaand Yasuda, 2005). For robots to succeed as

motivational trainers, teachers and playmates their physical appearance, interactivity and

social behaviour must match the users’ needs and meet the demands of an ever changing

educational context. This calls for well-established guidelines as to how such robots should be

designed and in what ways interacting with them can support learning.

Within the field of Persuasive Design (PD) it is argued that technologies can be persuasive in

the role as social actors. Naturally, from a PD perspective social robots can viewed as social

actors since their main purpose is to engage in social interactions with humans and thus one

might consider the principles of PD when designing social robots for persuasive purposes.

From a Human-Robot Interaction (HRI) perspective, though, the design of robots for

particular (often persuasive) purposes, has also received a great deal of attention in recent

years. Thus, this paper shows that a lot can be gained from looking at the taxonomies of HRI

when designing persuasive social actors. Furthermore, this paper argues that for robots to be

persuasive within an educational context, one must also consider the conditions of this

particular context of application. Specifically, this paper proposes the combination of PD,

HRI and Didactics as a framework for designing robots that motivate play and learning. The

notion of Persuasive Educational and Entertainment Robotics (PEERs) is introduced and

based on a case study on the implementation of the therapeutic robot seal PARO at a school

for children with autism, the different roles of the robot in persuasive interventions are

outlined and discussed in relation to the framework. The paper is concluded with thoughts on

the perspectives and limitations of the work with directions for future research.


Persuasive Educational and Entertainment Robotics (PEERs)

This paper builds on the work in (Bertel, 2012) which

argues that the combination of theory on motivation,

interaction and learning within PD, HRI and Didactics

can provide a framework for designing robots

specifically for the purpose of education. Thus, the

PEERs model (see figure 1) is a conceptualization of

robots designed to motivate play and learning as well

as a categorization of related fields and technologies;

A. Persuasive Robotics, i.e. robots that motivate

behavior change through social support (e.g.

robotic weight consultants that keep track of

users’ dietary information (Kidd & Breazeal,

2008) or service robots that elicits emotional

responses to users’ energy consumption

(Vossen, Ham & Midden, 2010) to encourage a healthier or greener lifestyle).

B. Educational Robotics, i.e. hands-on robotic kits such as LEGO Mindstorms used to

teach topics within STEM education. These robotic kits have generally been argued to

support a constructionist approach to learning (Papert, 1980) and thus facilitate

embodied learning through experimentation, reflection and collaboration (Mataríc,

Koenig & Feil-Seifer, 2007 and Majgaard & Misfeldt, 2011)

C. Persuasive Learning Designs (PLD), also termed Persuasive Learning Objects and

Technologies (EuroPLOT, 2010) covers the application of ICT tools to motivate and

enhance learning. Theoretical concepts and design methodologies have been

developed in (Gram-Hansen, 2012 and Gram-Hansen, Schärfe & Dinesen, 2012) with

a particular focus on including the context of the interaction in the design of PLD’s.

The conceptualization of PEERs creates the opportunity to compare these related research

fields theoretically and compile their respective strategies and principles when designing

social robots specifically with the purpose of motivating play and learning.

Robots in Persuasive Design and Persuasion in Robotics

The fields of HRI and PD have in common that they both argue robots as possible persuasive

agents. From an industrial HRI perspective, robots are most often considered to be tools (with

the primary purpose of increasing efficiency and reducing complex tasks and thus costs).

However, from a PD perspective this is not viewed as particularly persuasive even though it

quite possibly induces behavior (and perhaps even attitude) change. Within PD some

technologies described as simulated objects could also be considered robotic (e.g. the well-

known Baby Think It Over which provides compelling experiences to young women about

the advantages and (particularly) disadvantages of early motherhood (Fogg, 2003). So

according to the PD framework, a social robot that imitates the behavior of an animal or a

human convincingly can also be viewed as a portable persuasive simulation technology

provided that it seeks to highlight the impact of certain behaviors and motivate behavior or

attitude change (Fogg, 2003). However, although it seems that whether a robot should be

considered a tool (reducing complex tasks), a simulating medium (simulating physical,

psychological or emotional intelligence) or a social actor (engaging in social interaction,

providing feedback and creating social relations) would depend on the design of the

technology, the task at hand and the context of the interaction, most of the principles that can

be directly related to social robotics come under the persuasive role as a social actor. The

Figure 1. Persuasive Educational

and Entertainment Robotics


principles of this persuasive role are not limited to robots and include other technologies that

possess social qualities as well (virtual avatars etc.). Here, it is argued that the fact that people

respond socially to technologies has significant implications for persuasion as it allow these

technologies to apply a host of persuasion dynamics that arises from social situations (Fogg,

2003). These social cues are then formulated into a set of PD principles (see table 1) that

describes particular ways of utilizing these cues of social influence to motivate and persuade:

Table 1. Social Cues and Principles of Persuasive Social Actors

Cue Examples Persuasive Principle

Physical Face, eyes, body, movement Visually attractive PT is likely to be more persuasive

Psychological Preferences, humour,

personality, feelings, empathy

People are more readily persuaded by PT products

that are somewhat similar to themselves in some way

Language Interactive language use, spoken

language, language recognition

By offering praise (words, images, symbols, sounds)

PT can lead users to be more open to persuasion

Social Dynamics Turn taking, cooperation, praise,

answering questions, reciprocity

People will feel the need to reciprocate when PT has

done a favour for them

Social role Doctor, teammate, teacher, guide The role of authority enhances powers of persuasion

Within robotics, the research field of social robots emerged from the area of biologically

inspired robots and the idea that robots should be able to interact and share information, not

with humans, but among each other, as is the case with swarm robots (Dautenhahn & Billard,

1999). From the beginning of this millennium, though, social HRI gained increased interest,

and in 2003 the term Socially Interactive Robotics (SIR, i.e. robots whose sole purpose is to

engage in social interaction) was defined (Fong & Dautenhahn, 2003) and a taxonomy of

characteristics and key components of SIR developed (see table 2).

Table 2. Taxonomy of Socially Interactive Robots (SIR)

Properties Example Description

Morphology anthropomorphic, zoomorphic,

caricatured, functional

Establishes social expectations and provides

information about the intended function of the robot

Emotions anger, fear, sadness, joy, surprise,

neutral and combinations

Facilitate credibility in HRI and serve as feedback to

the user about the robots internal state

Dialogue synthetic language, natural

language, non-verbal cues

Exchange and interpretation of symbols and

information about the context of the interaction

Personality tool, pet, character, supernatural,

human-like

A set of qualities which are particularly significant

for a specific robot

Perception face/gaze tracking, speech/gesture

recognition, tone of voice

A social robot must possess a number of perceptual

abilities to engage in social interaction

User Modelling qualifications, experience,

cognitive abilities

The ability to create different user models so as to

adapt to and shape the interaction in relation to

specific user characteristics

Situated learning imitation, learning The ability to transfer information, skills and tasks

between robots and humans

Intentionality targeted behaviour, theory of

mind, joint attention

For people to be able to asses and predict behaviour,

it is necessary that the robot expresses intentionality


Some of the characteristics in this traditional HRI taxonomy are similar to the social cues

highlighted in PD. For instance the terms morphology and emotions cover qualities much like

the physical and psychological cues of persuasive social actors. Dialogue can be viewed as a

combination of the cues related to language and social dynamics. Finally, the robots

personality is also similar to PD’s idea about the social role of the technology. The remaining

components revolve around particular cognitive abilities (perception, user modeling, situated

learning and intentionality). One could say that the first four types of characteristics of HRI

similar to PD focus on qualities that shape the social interaction, whereas the last four types

characterize the prerequisites for fully autonomous social HRI.

The idea that robots can motivate behavior change through social support is not unknown to

the field of HRI. This is also referred to as the overlap between SIR and Assistive Robotics,

i.e. Socially Assistive Robotics (SAR), which most often describes assistive robotic

technologies that support physical or cognitive rehabilitation not by physical manipulation but

through social interaction (Feil-Seifer & Mataríc, 2005). The SAR framework (see table 3)

adds to the existing taxonomy for social robots the following concepts:

Table 3. Taxonomy of Socially Assistive Robots (SAR)

Properties Example Description

User

population

elderly, users with physical or

cognitive disorders, students

SAR can address various populations of users, ranging in age,

impairment, and need.

Task

tutoring, physical therapy,

daily life assistance, emotional

expression

SAR must engage the user effectively, achieve the goals of

the domain-specific activity and be responsive to the needs

and requirements of direct users and indirect stakeholders

Interaction

modality Speech, Gestures, Direct Input

SAR interactions vary in type and modality. Interactions

should be treated separately from the robot’s personality and

also describe the reciprocal interaction by the human user.

Role Caregiver, therapy aid,

companion

The role of the robot may be defined by the task it is assisting

with, the user population it is working with and the

impression it gives through its appearance and behaviour.

It can be argued, that these SAR concepts define Persuasive Robotics, i.e. the overlap between

PD and HRI (see figure 1) which is distinguished from other types of social robots since

having a predefined user group and task entails a specific (persuasive) intention. However, the

SAR taxonomy has much greater focus on the target user and the context of the interaction.

Positioning PEERs: The Importance of Context in Persuasive Learning Designs

As argued in (Gram-Hansen, Schärfe & Dinesen, 2012) the traditional approach to PD as a

context-independent framework limits its applicability as a design tool, particularly within the

field of education. This holds for the principles related to the persuasive role as social actor as

well. Here, the complex and highly contextual social cues that cause people to make

inferences about social presence (i.e. physical, psychological, language, social dynamics and

social roles) are reduced to a few narrow and unilateral principles revolving around only a

fragment of those social cues (i.e. visual attractiveness, similarity and reciprocity, praise and

authority). It could be argued that whether a social robot will be considered visually attractive

depends on the user and the nature and purpose of the interaction. For instance, a human-like

design was specifically avoided when designing a social robot to assist in the relocation of

patients (Mukai et al, 2010). Its size, colors and visual design were chosen to signal

cleanliness and robustness, but could also possibly be seen as intimidating in other contexts.


Similarly, as argued in (Bertel, 2010) the effects of praise and rewards are not constant over

time or across contexts. Rather, they are mediated by the power structure in the social context

of the interaction and the relationship between persuader and persuadee. Furthermore,

although a technology can certainly be persuasive in the role as an authority, it is not difficult

to imagine situations where authoritative behavior will have a detrimental effect on

motivation, particularly within education. Thus, when it comes to the design of robots as

social actors the PD principles do not really reflect the complexity or cultural and contextual

dependency of the aforementioned social cues or provide a method for prioritizing and

operationalizing them in specific designs. The taxonomic additions in SAR, however,

contribute to both the PD and HRI frameworks as it stresses the importance of context in

persuasive HRI and argues that particularly the role of the robot is defined by the task it is

assisting with, the user population it is working with and the impression it gives through its

appearance and behaviour (Feil-Seifer & Mataríc, 2005).

PEERs at Play: A Case Study in Autism Education

To understand in what ways PEERs can facilitate motivation and what roles they can take on

for specific educational purposes, we conducted a three month case study on the use of the

social robot PARO at a school for children with autism.

PARO is an autonomous robotic seal equipped

with sensors and actuators and computational

intelligence that enables it to simulate the sounds

and movements of a real baby harp seal. It is

being used with great success in dementia care in

Japan and Europe as a robot companion with the

purpose of increasing quality of life and reduce

stress and anxiety and to provide a therapeutic

tool for specific individual interventions (i.e.

calming down, keeping/removing focus, reviving

memories or language rejuvenating the identity as

a caregiver etc. (Klein, Gaedt & Cook, 2013 and

Wada & Shibata, 2007). Recent research suggests that PARO can be used as a facilitator of

social communication for children with autism as well (e.g. Marti et al., 2005 and Roberts &

Shore, 2013).

Whereas much of the related research on PARO for autism education were done in

experimental settings, in this case we did a 3 month field study on the implementation and use

of PARO at school for children with autism. This provided the opportunity to study long-term

interactions in real-world educational settings as well as to promote a high level of user

involvement in the development of the didactic designs supporting the use of PARO as a

PEER. A total of three PAROs were implemented in 3 groups of children divided by age but

with similar level of cognitive development (0-1 years of age) in which most have no spoken

language and use alternative tools for communication. In total, 20 children and 7 teachers

participated in study. The didactic applications of PARO were developed entirely by the

participating teachers through 3 participatory design phases; exploration, co-ideation and co-

creation/evaluation. The research setup and participatory framework is described in detail in

(Bertel, Rasmussen & Christiansen, 2013). The following analysis will thus focus on the role

of PARO as a PEER and the persuasive principles utilized in relation to this role, in particular.

Figure 2. The PEER PARO


The Role of PARO as a PEER in Persuasive Interventions

From the case study we found, that PARO had different roles in different teacher-developed

didactic designs depending on the individual child, the goal of the intervention and the

specific context of the interaction. The observed interventions were often centered on

facilitating specific types of attention (bodily/verbal/social) and can be categorized as follows:

Bodily/Verbal Attention (see figure 3)

PARO was used as a sensory/cognitive stimulant to motivate

verbalization and meaningful physical interaction (e.g. eye-

contact, petting) either independently or supported by a

teacher. This didactic design was deployed with a child, who

would mostly touch things very lightly with his fingertips.

With PARO he would engage in bodily interactions (using

arms, chest and face) for longer periods of time. Also, the children were encouraged to

verbalize their thoughts while interacting with PARO (e.g. “dog/ seal”, “feed/sleep” and “tail/

eyes”). The teachers reported that the children were particularly interested in the eyes and

would instantly seek eye-contact, which they would normally avoid. Here, it was thus the

morphology, i.e. the visual, tactile and auditory feedback as well as the simulated perception

(gaze-tracking, reaction to touch) that supported the persuasive intervention in question.

Joint Attention (see figure 4)

PARO was used as an object of joint attention between to

facilitate social interaction. E.g., one group of teachers used

the act of grooming PARO to get two children of whom one

was afraid of the other to jointly touch PARO and eventually

touch each other. They were already working on this

particular goal. However, the children had never shaken hands

or touched each other spontaneously until the sessions with

PARO. In other cases, PARO was used with three children in

the introduction and rehearsal of social concepts such as care giving (e.g. grooming, feeding,

and tucking in) and social events such as birthdays, dance parties and tobogganing. The

teacher explained how the one child with more ability to get ideas for social concepts would

instruct the second child who would then instruct the third child in the concepts he had

learned from the first child. Here, PAROs morphology, personality and perceptive abilities

motivated the children to take on the social role as teammates and sometimes even teachers.

Center of Attention (see figure 5)

PARO was used to create a centre of attention for a larger

group of children. The context would often be a social

gathering of some sort, e.g. singing Christmas Carols or

playing musical instruments. The children would gather in a

circle around a table or on the floor and PARO would be

placed in the centre. The interaction would be initiated and

maintained by the teacher and there would be less physical

interaction between PARO and the children. In some cases

PARO would even be turned off. However, it would be

included in the social setting of the interaction, often initiated by the children. One teacher

described how the children would provide PARO with a songbook of its own. Another

teacher described how a child had changed the lyrics from a lullaby to be about PARO.

Fig. 3. Interacting with

PARO to motivate bodily

or verbal attention

P C

(t)

(t)

P

C C


PARO to motivate joint

attention


PARO to motivate social

attention


Directing and redirecting attention (see figure 6)

Finally, in some didactic designs, PARO was used to

direct or redirect the attention of a child in certain

difficult situations. For instance, if a child would have

to stay focused on a specific task for longer than

usual (e.g. for psychological evaluation), or if a child

was intimidated either by external factors such as new

and unfamiliar surroundings (e.g. going to a new

playground or to church at Christmas) or by the fact

that he or she were to be the center of attention at a specific social event (e.g. visiting and

introducing oneself at a new school). In these cases PARO would be used strategically to shift

focus from the child’s activities and performance to the activities and performances of PARO

for which it would need the child’s ‘help’. Thus, rather than PARO accompanying the child,

the child would accompany PARO as a helper, teammate or teacher.

Conclusions and Directions for Future Work

Based on the empirical findings the role of the robot (as stressed in SAR) plays an important

part in defining the persuasive potential of social robots for persuasive, educational purposes.

In none of the above four types of teacher-developed didactic designs is there a particular

preference for the robot in the role as an authority, and especially in the latter a directly

inverse relationship between the user and the robot is considered the key to enhancing

motivation. That is, in this case the user is the authority (the teacher, guide or helper) and the

robot is the one requiring care and assistance. Thus, for the principles of PD (i.e.

attractiveness, similarity, authority, praise and reciprocity) to be useful as PEER design tools,

we argue that they must be extended and understood as the strategic use of particular

properties of HRI in relation to a specific target group, task and context within education. The

following is an attempt to do exactly this:

Table 4. Persuasive Principles of PEERs

Original

Principle

Extended/Revised

Principle

Related properties of

HRI Didactic Application

Attractiveness Strategic design of

physical cues

Morphology (SIR),

Personality (SIR)

PEERs physical appearance should be

attractive to both students and teachers

Similarity Strategic design of

psychological cues

Emotions (SIR),

Intentionality (SIR),

Perception (SIR)

PEERs should express intentions and

affective states meaningfully. If possible be

able to perceive these in users too

Praise and

rewards

Strategic

interaction design

Dialogue (SIR), Interaction

Modalities (SAR), Social

Dynamics (PD)

PEER interaction design must match the

physical appearance and user preference. If

possible it should be multimodal

Reciprocity Strategic design of

tasks

Task (SAR), Situated

learning (SIR)

PEERs should solve tasks not solvable by

humans. It should not replace humans.

Authority Strategic design of

social relations

Role (SAR), User

Modelling (SIR)

PEERs should be able to assume different

roles depending on user, task and context

In the future, these findings will be further explored in larger-scale, long-term and cross-

contextual case studies with morphologically different social robots, for different users, tasks

and contexts of interaction to investigate the applicability and replicability of these particular

persuasive principles of PEERs.


PARO to redirect attention


Acknowledgements

The authors would like to thank Egebakken and the participating teachers and students for

their contribution to this research. The research is done in cooperation with Aalborg

University and supported by the Danish Agency for Science, Technology and Innovation.

References

Bertel, L. B., Rasmussen, D. M. & Christiansen, E. (2013). Robots for Real: Developing a Participatory Design

Framework for Implementing Educational Robots in Real-world Learning Environments. INTERACT 2013

14th IFIP TC International Conference. Lecture Notes in Computer Science. In press

Bertel, L. B, (2013). Persuasive Educational and Entertainment Robotics (PEERs). In Bai, S., Larsen, J. A.,

Madsen, O. & Rehm, M (red). 1 st AAU Workshop on Human-Centered Robotics. Aalborg Universitetsforlag

Bertel, L. B. (2011). The Use of Rewards in Persuasive Design. Poster Proceedings of the 5th International

Conference on Persuasive Technology, Oulu University Press

Dautenhahn, K. & Billard, A. (1999). Bringing up robots or – the psychology of socially intelligent robots:

From theory to implementation. Proceedings of the Autonomous Agents

Dautenhahn, K. et al. (2009). KASPAR – a minimally expressive humanoid robot for Human-Robot Interaction

research (pp. 369-397). Applied Bionics and Biomechanics, 3 & 4

EuroPLOT (2010). Persuasive Learning Objects and Technologies. Retrieved June 2013, from http://eplot.eu

Feil-Seifer, D. & Mataríc, M. J. (2005). Defining Socially Assistive Robotics. 9th International Conference on

Rehabilitation Robotics.

Fogg, B. J. (2003). Persuasive Technology - Using Computers to Change What We Think and Do. Morgan

Kaufmann Publishers

Fong, T., Nourbakhsh, I. & Dautenhahn, K. (2003). A Survey of Socially Interactive Robots. Robotics and

Autonomous Systems, 42(3-4)

Gram-Hansen, S.B. (2012). PLOT Persuasive Learning Design Framework: Persuasive Learning Designs.

Deliverable D3.3 of the EU-project Persuasive Learning Objects and Technologies for Lifelong Learning in

Europe

Gram-Hansen, S. B., Schärfe, H. & Dinesen, J. V. (2012). Plotting to Persuade – Exploring the Theoretical

Cross Field between Persuasion and Learning. International Conference on Persuasive Technology

Han, J. & Dongho, K. (2009). R-Learning services for elementary school students with a teaching assistant

robot. 4th International Conference on Human-Robot Interaction

Kidd, C. D. & Breazeal, C. (2008). Robot at Home: Understanding Long-term Human Robot Interaction. IEEE

International Conference on Intelligent Robots and Systems

Klein, B., Gaedt, L. & Cook, G. (2013). Emotional robots: Principles and experiences with Paro in Denmark,

Germany and the UK. In GeroPsych: The Journal of Gerontopsychology and Geriatric Psychiatry, Vol 26(2)

Kozima, H., Nakagawa, C. & Yasuda, Y. (2005). Interactive Robots for Communication-care: A Case Study in

Autism Therapy. IEEE International Workshop on Robot and Human Interactive Communication, ROMAN

Lee, C. J., Kim, K., Breazeal, C. & Picard, R. (2008). Shybot: Friend-Stranger Interaction for Children Living

with Autism. CHI ’08 Extended abstracts on Human Factors in Computing Systems

Majgaard, G., Nielsen, J. & Misfeldt, M. (2010). Robot Technology and Numbers in the Classroom (pp.231-

234). CELDA 2010. International Association for Development

Marti, P., Pollini, A., Rullo, A. & Shibata, T. (2005). Engaging with artificial pets. Proceedings of Annual

Conference of the European Association of Cognitive Ergonomics

Mataríc, M. J., Koenig, N. & Feil-Seifer, D. (2007). Materials for enabling hands-on robotics and STEM

education. AAAI Spring Symposium, Resources for AI Education

Mukai, T. et al. (2010). Development of a Nursing-Care Assitant Robot RIBA That Can Lift a Human in Its

Arms. International Conference on Intelligent Robots and Systems.

Papert, S. (1980). Mindstorms: Children, Computers and Powerful Idea. Basic Books

Roberts, A. S. & Shore, S. M. (2013). Playing with a Robot: Enhancing Social Communication & Interaction.

Poster Presentation at International Meeting for Autism Research

Robins, B. et al. (2009). From Isolation to Communication: A Case Study Evaluation of Robot Assisted Play for

Children with Autism with a Minimally Expressive Humanoid Robot. 2nd IEEE International Conference on

Advances in Computer-Human Interactions

Vossen, S., Ham, J. & Midden, C. (2010). What Makes Social Feedback from a Robot Work? Disentangling the

Effect of Speech, Physical Appearance and Evaluation. International Conference on Persuasive Technology

Wada, K. & Shibata, T. (2007). Living with seal robots – its sociopsychological and physiological influences on

the elderly at a care house. IEEE Engineering in Medicine and Biology Magazine, 23, 5, 972-980


Proceedings of the International Workshop on EuroPLOT Persuasive Technology for Learning, Education...

Documents

Transcript of Proceedings of the International Workshop on EuroPLOT Persuasive Technology for Learning, Education...